System

Abstract

A system is provided.SOLUTION: A system comprising: means for receiving a voice input; means for transmitting received voice data to a server; means for converting the voice data into text data at the server; means for detecting a specific phrase in the converted text data; means for evaluating a risk level of the detected specific phrase; means for generating an alert based on an evaluation result; and means for notifying a pre-registered contact of the generated alert.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The technology of the present disclosure relates to a system. [Background technology]

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] In recent years, the methods of special fraud have become more diverse, and the amount of damage caused by them has been steadily increasing. Elderly people living alone are particularly likely to be targeted, and it is difficult to prevent fraud using conventional methods. Given this background, it is necessary to provide an effective system to prevent ordinary users, including the elderly, from falling victim to special fraud and to prevent fraud damage before it occurs. [Means for solving the problem]

[0005] The present invention provides a system for preventing special frauds, including a means for receiving voice input, a means for transmitting the received voice data to a server, a means for converting the voice data to text data at the server, a means for detecting specific phrases in the converted text data, a means for evaluating the risk level of the detected specific phrases, a means for generating an alert based on the evaluation results, and a means for notifying pre-registered contacts of the generated alert. This system analyzes the content of calls in real time and immediately sends an alert if there is a possibility of fraud, significantly reducing the risk of elderly people and general users falling victim to fraud. Furthermore, by including a means for analyzing received voice data and constantly learning the latest criminal techniques, and a means for attaching a call summary and recording when generating an alert, even more effective defense measures can be provided.

[0006] The "means for receiving voice input" refers to a device and method for capturing the user's voice during a call using a microphone or other voice input device.

[0007] The "means for transmitting received voice data to a server" refers to a communication device and protocol for transmitting the acquired voice data to a server via a network.

[0008] The "means for converting voice data into text data at the server" refers to software and algorithms that convert voice data received at the server into text using voice recognition technology.

[0009] "Means for detecting specific phrases within the converted text data" refers to software and algorithms that search for and extract predefined potentially fraudulent phrases or keywords from the text data.

[0010] "Means for assessing the riskiness of specific detected phrases" refers to software and evaluation criteria for scoring the likelihood of a detected phrase being fraudulent.

[0011] The "means for generating an alert based on the evaluation results" refers to software and procedures that generate warning information (alerts) based on the results of the risk evaluation.

[0012] "Means for notifying pre-registered contacts of generated alerts" refers to communications devices and protocols that send generated alert information to pre-registered contacts, such as the user's relatives or local police.

[0013] "Means for analyzing voice data and constantly learning the latest criminal methods" refers to software and algorithms that utilize machine learning technology using received voice data to continuously learn new fraud methods and incorporate them into the system.

[0014] "Call summary and recording attachment means" means software and procedures that attach call summary information and corresponding audio recordings to generated alert information. [Brief explanation of the drawings]

[0015] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8]FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0016] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0017] First, the terms used in the following description will be explained.

[0018] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0019] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0020] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0021] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0022] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0023] [First embodiment]

[0024] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0025] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0026] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0027] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0028] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0029] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0030] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0031] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0032] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0033] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0034] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0035] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0036] This invention is a system for preventing special frauds. The system consists of terminal, server, and user components that work together to monitor and evaluate potentially fraudulent calls in real time, and alert relatives and relevant authorities as necessary.

[0037] Receiving and sending audio input

[0038] The device collects the call audio in real time from the moment the user answers the call. Audio input is captured using a microphone or similar audio input device, and the audio data is then transmitted to a server, typically using WebSocket or another real-time communication protocol.

[0039] Speech recognition and text conversion

[0040] The server receives the voice data sent from the device. The received voice data is converted into text data by a voice recognition engine. This voice recognition engine uses the latest machine learning technology, enabling highly accurate voice-to-text conversion.

[0041] Deceptive word detection and evaluation

[0042] The server analyzes the text data generated by the speech recognition to detect specific fraud words and phrases, typically using a predefined list of fraud-related keywords, which are then fed into a scoring algorithm to assess risk.

[0043] Alerting and Notifications

[0044] Based on the results of the risk assessment, the server generates an alert, which includes a summary of the call, detected scam words, and the text data of the call. The alert is then sent to pre-registered contacts, such as the user's relatives or relevant authorities. Notifications can be sent via email, SMS, in-app notifications, and other methods.

[0045] Always responding to the latest criminal techniques

[0046] The server continuously analyzes the received audio data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[0047] Specific examples

[0048] For example, if a user uses a phrase such as "Please hurry up and transfer the money" during a call, the device sends the voice data to the server in real time. The server converts the voice data into text and detects the phrase "Please hurry up and transfer the money." The risk scoring algorithm evaluates this phrase as high risk and immediately generates an alert to notify relatives and the police. This process can prevent fraud before it occurs.

[0049] This invention can effectively prevent damage caused by special frauds, and can significantly reduce the risk of being defrauded, especially for elderly people and single people. The specific programs and algorithms involved in implementing the invention will vary depending on the technology and environment to which it is applied, but they will generally be implemented based on the form shown here.

[0050] The processing flow will be explained below.

[0051] Step 1:

[0052] The terminal detects the call start event as soon as the user answers the phone and starts voice input. The voice is picked up in real time through a microphone and temporarily stored in a buffer as a digital signal.

[0053] Step 2:

[0054] The device transmits the captured audio data to the server in real time using WebSocket or other real-time communication protocols.

[0055] Step 3:

[0056] The server receives the voice data sent from the terminal, converts the received voice data into an appropriate format, and then inputs it into a voice recognition engine.

[0057] Step 4:

[0058] The server's speech recognition engine converts the speech data into text data. This engine performs highly accurate conversion from speech to text, and the generated text data is used for analysis.

[0059] Step 5:

[0060] The server analyzes the converted text data to detect specific fraudulent words and phrases using a pre-trained list of fraud-related keywords.

[0061] Step 6:

[0062] The server uses a scoring algorithm to assess the risk of detected scam words, with different weights assigned to different scam words to quantify their risk.

[0063] Step 7:

[0064] The server generates an alert based on the results of the scoring algorithm, which includes a summary of the call, detected scam words, reputation score, and text data.

[0065] Step 8:

[0066] The server then sends the generated alerts to pre-registered contacts of relatives and relevant authorities via email, SMS, in-app notifications, and more.

[0067] Step 9:

[0068] The server feeds the received voice and text data into a machine learning algorithm to constantly learn the latest fraud techniques, and the results of this learning are used for subsequent analyses.

[0069] Step 10:

[0070] The server stores all transaction logs in a database for future reference, including the start and end times of calls, analysis results, and notification content.

[0071] By going through the above steps, the present invention functions as an effective system for protecting users from special fraud.

[0072] Example 1

[0073] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0074] Currently, the methods of special fraud are evolving daily, increasing the risk of elderly people and single people falling victim to fraud. In particular, methods such as "bank transfer fraud" carried out over the phone are becoming more sophisticated, making it difficult for conventional prevention systems to deal with them. Therefore, a system is needed that analyzes the content of phone calls in real time, evaluates the risk of fraud, and issues alerts.

[0075] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0076] In this invention, the server includes a means for receiving voice input, a means for transmitting the received voice data to the server using a real-time communication protocol, and a means for converting the voice data into text data at the server, thereby enabling real-time analysis of the user's call content, detection of specific fraudulent words and phrases, and highly accurate risk assessment and prompt alert notification.

[0077] "Voice input" refers to a technology that uses a microphone or other voice input device to collect voices uttered by a user via a communication means such as a telephone.

[0078] A "real-time communication protocol" is a communication protocol for transmitting data such as telephone voice promptly with low latency. An example is WebSocket.

[0079] A "server" is a computer system that receives voice data sent from a terminal and performs processes such as voice recognition and data analysis.

[0080] "Text data" refers to character string information generated from voice data using voice recognition technology.

[0081] "Fraudulent words and phrases" are specific, predefined words or phrases related to fraudulent activity, such as "transfer" or "hurry."

[0082] A "scoring algorithm" is a calculation method used to numerically assess the risk of detected fraudulent words or phrases.

[0083] An "alert" is a warning message that is generated when dangerous call content is detected.

[0084] "Contacts" are people or organizations to whom alerts are sent. They are selected based on pre-registered information.

[0085] "Email" refers to digital letters sent and received over the Internet.

[0086] "SMS" is a means of sending and receiving short text messages over mobile phone networks.

[0087] "In-app notifications" refer to notification messages displayed through software such as smartphone apps.

[0088] The present invention is a system for preventing special frauds in advance, in which the components of the user, terminal, and server work together.

[0089] Receiving and sending audio input

[0090] The device collects the call audio in real time using a voice input device (e.g., a smartphone microphone) from the moment the user answers the call. The audio data is then sent to the server using a real-time communication protocol such as WebSocket.

[0091] Speech recognition and text conversion

[0092] The server receives the voice data sent from the device. The received voice data is converted into text data by a speech recognition engine (e.g., Google Cloud Speech-to-Text API). This speech recognition engine uses the latest machine learning technology, enabling highly accurate speech-to-text conversion.

[0093] Deceptive word detection and evaluation

[0094] The server analyzes the text data generated by the speech recognition. Natural Language Processing (NLP) algorithms are used for the analysis. The server uses a predefined list of fraud-related keywords to detect specific fraud words and phrases within the text data. The detected fraud words are then evaluated for risk using a scoring algorithm.

[0095] Alerting and Notifications

[0096] The server generates an alert based on the results of the risk assessment. The alert includes a summary of the call, detected scam words, and text data. The alert is then sent to pre-registered contacts, such as the user's relatives or relevant authorities, via email, SMS, or in-app notification.

[0097] Always responding to the latest criminal techniques

[0098] The server continuously analyzes the received audio data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[0099] Specific examples

[0100] For example, if a user uses a phrase such as "Please transfer the money quickly" during a call, the device will send the speech data to the server in real time. The server will convert the speech into text and detect the phrase "Please transfer the money quickly." The risk scoring algorithm will evaluate the phrase as high risk and immediately generate an alert to notify relatives and the police. This process can prevent fraud before it can be carried out.

[0101] Examples of prompt statements

[0102] "Detect specific phrases in the audio of a call and assess the likelihood of fraud. For example, if the phrase 'Please transfer the money quickly' is detected, issue a warning."

[0103] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0104] Step 1:

[0105] The device starts voice input.

[0106] The device uses a microphone to collect the call voice in real time from the moment the user answers the call. Specifically, the voice input software on the device runs in the background and collects the user's voice data.

[0107] Input: User's voice

[0108] Output: Collected audio data

[0109] Step 2:

[0110] Sending audio data to the server

[0111] The device sends the collected audio data to a server using a real-time communication protocol such as WebSocket, which requires stable communication with low latency.

[0112] Input: Collected audio data

[0113] Output: Audio data sent to the server

[0114] Step 3:

[0115] The server receives the audio data.

[0116] The server receives the voice data sent from the terminal in real time. The server process stores the received data and prepares it for subsequent voice recognition processing.

[0117] Input: Audio data sent from the device

[0118] Output: Archived audio data

[0119] Step 4:

[0120] Converting audio data to text

[0121] The server converts the voice data into text using a speech recognition engine, for example, the Google Cloud Speech-to-Text API, which ensures high accuracy in the process.

[0122] Input: Archived audio data

[0123] Output: Text data

[0124] Step 5:

[0125] Text data analysis

[0126] The server analyzes the text data generated by the speech recognition, using natural language processing (NLP) algorithms to detect fraudulent words and phrases, using a predefined list of fraudulent words.

[0127] Input: Text data

[0128] Output: Detected fraudulent words and phrases

[0129] Step 6:

[0130] Risk assessment

[0131] The server evaluates the detected fraudulent words and phrases using a scoring algorithm, which generates a numerical score indicating the likelihood of fraud.

[0132] Input: Detected scam words and phrases

[0133] Output: Risk score

[0134] Step 7:

[0135] Generate alerts

[0136] The server generates an alert based on the results of the risk assessment, which includes a summary of the call, detected scam words, and text data.

[0137] Input: Risk score, text data

[0138] Output: Alert

[0139] Step 8:

[0140] Alert Notification

[0141] Generated alerts are sent to pre-registered contacts via email, SMS, or in-app notifications.

[0142] Input: Alert

[0143] Output: Notified relatives and relevant authorities

[0144] Step 9:

[0145] Learning new fraud methods

[0146] The server uses the received voice data and analysis results to automatically learn new fraudulent techniques through machine learning algorithms, allowing the system to adapt to the latest fraud techniques.

[0147] Input: Received voice data, analysis results

[0148] Output: Updated deceptive wordlist and scoring algorithm

[0149] (Application example 1)

[0150] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0151] In modern society, special frauds are on the rise, with victims becoming increasingly serious, especially among the elderly. Telephone frauds are particularly sophisticated, leaving many victims vulnerable to deception. Current technology does not provide sufficient means to detect and prevent these frauds in real time. Therefore, a system is needed that analyzes the content of phone calls in real time and immediately sends an alert if there is a possibility of fraud. The system itself must also be able to learn and evolve so that it can constantly respond to the latest fraud techniques.

[0152] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0153] In this invention, the server includes a means for receiving voice input, a means for transmitting the received voice data to the server, a means for converting the voice data to text data at the server, a means for detecting specific phrases in the converted text data, a means for evaluating the risk level of the detected specific phrases, a means for generating an alert based on the evaluation results, a means for notifying pre-registered contacts of the generated alert, a means for collecting voice data and detecting fraudulent keywords in real time, and a means for automatically generating an alert to notify contacts when fraudulent keywords are detected. This allows the content of a user's calls to be analyzed in real time, and an alert to be immediately generated if there is a high possibility of fraud, making it possible to prevent damage before it occurs. Furthermore, by analyzing received voice data and constantly learning the latest criminal techniques, the system continuously evolves and can respond to more advanced fraud techniques.

[0154] "Audio input" refers to the means by which an audio signal is input to a digital device.

[0155] "Device" means an electronic device used by an individual or organization for communication and data processing.

[0156] "Server" refers to a computer system that receives, stores, analyzes, and transmits data over a network.

[0157] "Audio data" refers to data that represents an audio signal in digital form.

[0158] "Text data" refers to data obtained by converting voice data into character information.

[0159] "Specific phrases" refer to specific words or expressions that indicate possible fraud.

[0160] "Real-time" refers to data collection, analysis, notification, etc. being carried out immediately without delay.

[0161] "Risk assessment" refers to the process of analyzing and assessing the risk level of specific detected phrases.

[0162] An "alert" is a message that sends a warning or notification when a specific condition is met.

[0163] "Pre-registered Contacts" refers to contact information that has been pre-registered in the system for the purpose of sending alerts.

[0164] "Scam words" refer to specific keywords or phrases that indicate potential fraud.

[0165] "Machine learning" refers to technologies and methods that allow computers to analyze large amounts of data and learn automatically through algorithms.

[0166] The system of the present invention is designed to prevent special frauds, and involves the cooperation of the terminal, server, and user components. First, when a user starts a call, the terminal collects the user's voice input in real time and transmits it to the server as voice data. This voice data is transmitted using WebSocket or other real-time communication protocols.

[0167] The server converts the received voice data into text data using a speech recognition engine. This speech recognition engine uses the latest machine learning technology, such as Google's Speech-to-Text API, to achieve highly accurate speech-to-text conversion. The converted text data is then sent to the scam word detection and evaluation process.

[0168] To detect fraudulent words, a predefined list of fraud-related keywords is used. The server analyzes the generated text data to check whether it contains any of the keywords in this list. If any of these keywords are detected, a scoring algorithm is applied to assess the risk level. This algorithm evaluates the risk level of the detected keywords and generates an alert based on the results.

[0169] The generated alert will be sent to pre-registered contacts via email, SMS, or in-app notification, including a summary of the call and a recording. The alert will include information on the call that is likely to be fraudulent, allowing the user's relatives or relevant authorities to take immediate action. Notifications can be sent via notification services such as Amazon SNS.

[0170] Furthermore, the system uses machine learning algorithms to continuously analyze the received voice data in order to constantly adapt to the latest criminal techniques. This allows the system to respond quickly to new fraud methods. When a user actually experiences a call containing fraudulent words, the system accumulates that experience as learning data to improve future detection accuracy.

[0171] For example, if a user makes a call containing the phrase "Please hurry up and transfer the money," the device will send the voice data to the server in real time. The server will convert the voice data into text and detect the phrase "Please hurry up and transfer the money." Based on this detection result, a risk assessment will be performed and if the call is deemed high risk, an alert will be generated immediately and notified to relatives and relevant authorities. This process can prevent fraud from being carried out.

[0172] An example of a prompt sentence is, "There is a risk of important information being leaked during the phone call. Please evaluate the phrase 'Please transfer the money quickly' as a possible fraud and generate an alert."

[0173] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0174] Step 1:

[0175] When a user receives a call, the device collects the voice input through a microphone. This voice input is recorded as digital voice data. The input data is processed in real time and then sent to the next step. The output is the captured voice data.

[0176] Step 2:

[0177] The device sends the collected audio data to the server over the network. This process uses WebSocket or other real-time communication protocols. The input is the audio data collected on the device, and the output is the audio data sent to the server.

[0178] Step 3:

[0179] The server converts the received voice data into text data using a voice recognition engine, such as Google's Speech-to-Text API. The input is the voice data sent to the server, and the output is the converted text data.

[0180] Step 4:

[0181] The server analyzes the generated text data and detects fraudulent words using a predefined list of fraud-related keywords. It checks whether the text data contains any of the keywords in this list. The input is the text data, and the output is a list of detected fraudulent words.

[0182] Step 5:

[0183] The server applies a scoring algorithm to assess the risk of detected fraudulent words. The algorithm assesses the risk level of the detected keywords and outputs the result. The input is a list of fraudulent words, and the output is a risk assessment score.

[0184] Step 6:

[0185] Based on the assessment results, the server generates an alert, which is accompanied by information such as a summary of the call and the recording. The input is the risk assessment score and text data, and the output is the alert message.

[0186] Step 7:

[0187] Generated alerts are sent to pre-registered contacts via email, SMS, or in-app notification. This is done by the server using a notification protocol. The input is the alert message and the output is the notified alert.

[0188] Step 8:

[0189] To stay up to date with the latest criminal techniques, the server continuously analyzes the received audio data and uses machine learning algorithms to learn new fraud techniques. The input is previously processed audio data, and the output is an updated fraud wordlist and rating model.

[0190] This can prevent users from being involved in fraudulent activities and improve the accuracy of the system.

[0191] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0192] This invention combines an emotion engine with a system for preventing special frauds. The system's components, including the terminal, server, and user, work together to monitor and evaluate potentially fraudulent calls in real time, alerting relatives and relevant authorities as necessary. It also simultaneously recognizes the user's emotional state and incorporates that information into the evaluation, enabling more accurate fraud prevention.

[0193] Receiving and sending audio input

[0194] The device collects the call audio in real time from the moment the user answers the call. Audio input is captured using a microphone or similar audio input device, and the audio data is then transmitted to a server, typically using WebSocket or another real-time communication protocol.

[0195] Speech recognition and text conversion

[0196] The server receives the voice data sent from the device. The received voice data is first converted into an appropriate format and then input into a voice recognition engine. The voice recognition engine converts the voice data into text data. This engine performs highly accurate conversion from voice to text and the generated text data is used for analysis.

[0197] Deceptive word detection and evaluation

[0198] The server analyzes the text data generated by the speech recognition to detect specific fraud words and phrases using a predefined list of fraud-related keywords, which are then fed into a scoring algorithm to assess risk.

[0199] Emotion recognition by emotion engine

[0200] The server then inputs the received voice data into an emotion engine to analyze the user's emotional state in real time. The emotion engine analyzes changes in voice tone, speed, and volume to identify emotions such as joy, anger, sadness, and anxiety.

[0201] Evaluations that reflect emotional states

[0202] The server then evaluates the detected fraudulent phrases along with the user's emotional state as recognized by the emotion engine. If the user is in a certain emotional state (e.g., anxious or confused), points are added to the score to increase the likelihood of fraud.

[0203] Alerting and Notifications

[0204] Based on the results of the risk assessment, the server generates an alert, which includes a summary of the call, detected scam words, reputation score, text data, and the perceived emotional state. The alert is then sent to the user's contacts, including their next of kin and relevant authorities. Notifications can be sent via email, SMS, in-app notifications, and other methods.

[0205] Always responding to the latest criminal techniques

[0206] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[0207] Specific examples

[0208] For example, if a user uses a phrase like "Please hurry up and transfer the money" during a call, while the emotion engine detects an emotional state such as anxiety or confusion, the device sends the voice data to the server in real time. The server converts the voice into text and detects the phrase "Please hurry up and transfer the money" and the user's sense of anxiety. The risk scoring algorithm evaluates these as high risk and immediately generates an alert to notify relatives and the police. This process can prevent fraud from being committed.

[0209] This invention can effectively prevent damage caused by special frauds, and can significantly reduce the risk of being defrauded, especially for elderly people and single people. The specific programs and algorithms involved in implementing the invention will vary depending on the technology and environment to which it is applied, but they will generally be implemented based on the form shown here.

[0210] The processing flow will be explained below.

[0211] Step 1:

[0212] When the user answers the phone, the terminal detects the call start event and starts voice input. The call voice is picked up through the microphone and stored in a buffer in real time as a digital signal.

[0213] Step 2:

[0214] The device transmits the captured audio data to the server in real time, typically using WebSocket or other real-time communication protocols.

[0215] Step 3:

[0216] The server receives the voice data sent from the device and converts it into an appropriate format, which is then sent to a voice recognition engine.

[0217] Step 4:

[0218] The server's speech recognition engine converts the speech data into text data, which is then analyzed.

[0219] Step 5:

[0220] The server analyzes the generated text data to detect specific fraudulent words and phrases using a predefined list of fraud-related keywords.

[0221] Step 6:

[0222] The server uses a scoring algorithm to assess the risk of detected fraudulent words, which assigns different weights to different fraudulent words and quantifies their risk.

[0223] Step 7:

[0224] In parallel, the server inputs the voice data into an emotion engine to analyze the user's emotional state in real time, analyzing parameters such as voice tone, speed, and volume to identify emotions.

[0225] Step 8:

[0226] The server evaluates the risk level by taking into account the detected fraudulent phrases and the user's emotional state as recognized by the emotion engine. For example, if the user is anxious or confused, an evaluation score is added.

[0227] Step 9:

[0228] Based on the scoring algorithm, the server generates an alert that includes a summary of the call, detected scam words, reputation score, text data, and the perceived emotional state.

[0229] Step 10:

[0230] The generated alerts are sent by the server to pre-registered contacts via email, SMS, in-app notifications, and other methods.

[0231] Step 11:

[0232] The server feeds the received voice and text data into a machine learning algorithm to constantly learn the latest fraud techniques, allowing it to respond to new fraud methods.

[0233] Step 12:

[0234] The server stores all processing logs in a database, including the start and end times of calls, analysis results, and notification content.

[0235] By going through these steps, the system can protect users from special fraud and provide a fast and effective response.

[0236] Example 2

[0237] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0238] Special fraud methods are becoming more sophisticated every year, and the elderly and people living alone are at particularly high risk of falling victim to fraud. Conventional fraud prevention systems have the problem of being unable to completely prevent fraud victims because they have difficulty accurately identifying the content of fraud and are unable to reflect the user's emotional state in their evaluation.

[0239] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for receiving voice input, means for transmitting the received voice data to the server, means for converting the voice data into text data at the server, means for detecting specific phrases in the converted text data, means for evaluating the risk level of the detected specific phrases, means for analyzing the received voice data and recognizing the user's emotional state in real time, means for reflecting the user's emotional state in the evaluation, means for generating an alert based on the evaluation result, and means for notifying the generated alert to a pre-registered contact. This enables the possibility of fraud to be detected with high accuracy and enables a prompt and accurate response that also takes into account the user's emotional state.

[0240] "Means for receiving audio input" refers to a microphone or similar audio capture device for capturing audio data, such as speech.

[0241] "Means for transmitting received voice data to a server" refers to a method for transferring collected voice data to a server via a communication network such as the Internet.

[0242] "Means for converting voice data into text data at the server" refers to the process of converting voice data into text format using a voice recognition engine.

[0243] "Means for detecting specific phrases within the converted text data" refers to algorithms that search and detect pre-defined fraud-related keywords and phrases within the text data.

[0244] "Means for assessing the risk of specific detected phrases" refers to a method for quantifying and assessing the likelihood of fraud or the risk based on the detected phrases.

[0245] "Means for analyzing received voice data and recognizing the user's emotional state in real time" refers to an emotion engine that analyzes voice data and identifies the user's emotional state in real time, such as anxiety, anger, joy, or sadness.

[0246] "Means for reflecting the user's emotional state in the assessment" refers to a method for integrating and reflecting the user's emotional state recognized by the emotion engine in the fraud risk assessment.

[0247] "Means for generating alerts based on assessment results" refers to the process for generating warning messages or notifications when a high likelihood of fraud is assessed.

[0248] "Means of notifying pre-registered contacts of generated alerts" refers to means of sending generated alerts to pre-designated contacts, such as next of kin or relevant authorities, via email, SMS, in-app notifications, etc.

[0249] The present invention combines an emotion engine with a system for preventing special frauds, and is realized by the cooperation of the terminal, server, and user components.

[0250] Receiving and sending audio input

[0251] The device detects the start of a call when the user makes or receives a call. It catches the incoming call event and recognizes the beginning of the call. When the user starts a call, the device collects audio data of the conversation in real time using the built-in microphone or an external audio capture device. This audio data is sent to the server using a real-time communication technology such as the WebSocket protocol. To minimize communication delays, the audio data is sent intermittently, often using buffering.

[0252] Speech recognition and text conversion

[0253] The server receives the voice data sent from the device. The received voice data is stored in a specific buffer, where noise removal and voice normalization are performed as a preparatory step for analysis. The processed voice data is then input into a speech recognition engine (e.g., Google Speech-to-Text or IBM Watson) to convert the voice data into highly accurate text data. This text data is then analyzed.

[0254] Deceptive word detection and evaluation

[0255] The server analyzes the text data generated by the speech recognition engine. The analysis tool searches for and detects predefined fraud-related keywords and phrases (e.g., "transfer" or "hurry") within this text data. The detected fraud words are then input into a risk scoring algorithm, which evaluates the likelihood of the fraud.

[0256] Emotion recognition by emotion engine

[0257] In parallel, the server inputs the received voice data into the emotion engine to analyze the user's emotional state in real time. The emotion engine analyzes changes in voice tone, speed, and volume to identify the user's emotional state (e.g., anger, sadness, joy, anxiety, etc.).

[0258] Evaluations that reflect emotional states

[0259] The server integrates the detected fraud phrases with the user's emotional state as recognized by the emotion engine to evaluate the overall likelihood of fraud. In particular, if the user is anxious or confused, these feelings are added to the evaluation score.

[0260] Alerting and Notifications

[0261] Based on the evaluation results, the server generates an alert, which includes a summary of the call, detected scam words, evaluation score, text data, and the recognized emotional state, and the alert is sent to the user's relatives and relevant authorities via email, SMS, in-app notification, etc.

[0262] Always responding to the latest criminal techniques

[0263] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraud tactics, constantly updating the system's fraud detection capabilities.

[0264] Specific examples

[0265] For example, if a user says "Please hurry up and transfer the money" during a call, and the emotion engine detects the user's anxiety, the device will send the voice data to the server in real time. The server will convert the voice data into text, detect the phrase "Please hurry up and transfer the money," and the emotion engine will confirm the user's anxiety. The risk scoring algorithm will evaluate these as high risk and immediately generate an alert to notify relatives and the police. This process can prevent fraud from being committed.

[0266] Prompt Sentence Examples

[0267] Example prompts for input to a generative AI model:

[0268] "Please explain in detail the risk assessment process when the phrase 'hurry and transfer money' and the user's state of anxiety are detected during a call."

[0269] "Please tell me specifically how to send real-time collected call audio to a fraud detection system."

[0270] This system can significantly reduce the risk of elderly people and single people falling victim to fraud.

[0271] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0272] Step 1:

[0273] A call begins when a user makes or receives a call. The terminal catches this call reception event and recognizes the beginning of the call. The input includes the call reception event by the user. The output is the recognition of the start of the call.

[0274] Step 2:

[0275] When a user initiates a call, the device collects real-time audio data from the conversation using a built-in microphone or an external audio capture device. The input includes the user's voice during the call. The output is the collected audio data.

[0276] Step 3:

[0277] The device transmits the collected audio data to the server using real-time communication technology such as the WebSocket protocol. To minimize communication delays, the audio data is transmitted intermittently using buffering. The input includes the collected audio data. The output is the audio data transmitted to the server.

[0278] Step 4:

[0279] The server receives the audio data sent from the device. The received audio data is stored in a specific buffer, and noise removal and audio normalization are performed as a pre-analysis step. The input includes the audio data sent from the device. The output is the pre-processed audio data.

[0280] Step 5:

[0281] The server inputs the processed speech data into a speech recognition engine (e.g., Google Speech-to-Text or IBM Watson) to convert the speech data into text format. The input includes the preprocessed speech data. The output is the generated text data.

[0282] Step 6:

[0283] The server receives the generated text data from the speech recognition engine and inputs it into an analysis tool to detect predefined fraud-related keywords and phrases. The input includes the generated text data. The output is the detected fraud words and phrases.

[0284] Step 7:

[0285] The server inputs the detected fraud words into a risk scoring algorithm to quantify the likelihood of fraud. The input includes the detected fraud words or phrases. The output is a risk score.

[0286] Step 8:

[0287] In parallel, the server inputs the collected voice data into the emotion engine to analyze the user's emotional state in real time. The input includes the collected voice data. The output is the identification result of the user's emotional state.

[0288] Step 9:

[0289] The server integrates the detected scam phrases with the user's emotional state as recognized by the emotion engine to assess the overall likelihood of fraud. If the user is anxious or confused, this assessment is taken into account. Inputs include scam words, risk scores, and the user's emotional state. The output is the integrated assessment result.

[0290] Step 10:

[0291] Based on the evaluation results, the server generates an alert, which includes a summary of the call, detected scam words, the evaluation score, text data, and the recognized emotional state. The input includes the integrated evaluation results. The output is the generated alert.

[0292] Step 11:

[0293] The server notifies the generated alert to the user's relatives and relevant authorities via email, SMS, in-app notification, etc. The input includes the generated alert. The output is the notification sent.

[0294] Step 12:

[0295] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraud techniques, thereby constantly updating the system's fraud detection capabilities. Inputs include stored voice and text data. Outputs are the results of learning new criminal techniques.

[0296] (Application example 2)

[0297] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0298] Special frauds are becoming more sophisticated every year, and many people, especially the elderly, are falling victim to them. Frauds committed during phone calls have limited reaction time, making it difficult to respond immediately. Conventional fraud prevention systems detect fraud based on voice data, but do not take into account the user's psychological state, making it difficult to make accurate judgments. Since a rapid response to new fraud methods is also necessary, the present invention provides a system that solves these problems.

[0299] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving voice input, means for transmitting the received voice data to the server, means for converting the voice data to text data at the server, means for detecting specific phrases in the converted text data, means for evaluating the risk of the detected specific phrases, means for generating an alert based on the evaluation result, means for notifying pre-registered contacts of the generated alert, means for analyzing the user's emotional state, and means for reflecting the analyzed emotional state in a fraud risk evaluation. This enables highly accurate fraud detection and rapid response that takes the user's psychological state into consideration.

[0300] A "means for receiving voice input" is a device or system that converts a user's speech into an electrical signal and receives it in digital form.

[0301] The "means for transmitting received voice data to a server" refers to a method or protocol for transmitting voice data acquired by a terminal to a server via the Internet or other networks.

[0302] "Means for converting voice data into text data at the server" refers to the technology or algorithm that analyzes received voice data and converts the content into text format.

[0303] The "means for detecting specific phrases within the converted text data" refers to a filtering technique for identifying predefined specific words or phrases from the text data.

[0304] "Means for assessing the risk level of specific detected phrases" refers to algorithms or methods that assess the risk of the content of the call based on identified phrases and score them according to certain criteria.

[0305] The "means for generating alerts based on the evaluation results" is a system for creating warning messages and notifications based on the results of the risk evaluation.

[0306] "Means for notifying pre-registered contacts of generated alerts" refers to a mechanism for distributing generated warning messages to pre-defined contacts of next of kin and relevant authorities.

[0307] "Means for analyzing the user's emotional state" refers to technology that analyzes input data such as voice data and identifies the emotions (joy, anger, sadness, anxiety, etc.) felt by the user.

[0308] The "means for reflecting the analyzed emotional state in the fraud risk assessment" is a method for incorporating the user's emotional state into the risk assessment process and adjusting the risk score based on the results.

[0309] This invention combines an emotion engine with a system for preventing special frauds. Specifically, the system monitors users' calls in real time, assesses the possibility of fraud, and alerts relatives and relevant authorities if necessary.

[0310] The system is configured as follows:

[0311] Hardware and Software

[0312] The system is built using the user's smartphone, a server, and related software.

[0313] Hardware: Smartphone microphone, cellular network

[0314] Software: WebSocket (real-time transmission of voice data), requests (transmission of HTTP requests), speech-to-text API (conversion of voice data into text), emotion_recognition (analysis of emotional states), notification-service API (sending alerts)

[0315] Data processing and calculation

[0316] 1. Receiving and sending voice input

[0317] From the moment the user answers the call, the smartphone device collects the call audio in real time. Voice input is captured using a microphone, and then the audio data is sent to the server using WebSocket.

[0318] 2. Speech Recognition and Text Conversion

[0319] The server receives the voice data sent from the device. The received voice data is first converted into an appropriate format, and then converted into text data using a speech recognition engine. High-precision conversion from voice to text is performed, and the generated text data is then analyzed.

[0320] 3. Fraudulent word detection and evaluation

[0321] The server analyzes the text data generated by the speech recognition to detect specific fraud words and phrases using a predefined list of fraud-related keywords, which are then fed into a scoring algorithm to assess risk.

[0322] 4. Emotion Recognition by Emotion Engine

[0323] The server then inputs the received voice data into an emotion engine to analyze the user's emotional state in real time. The emotion engine analyzes changes in voice tone, speed, and volume to identify emotions such as joy, anger, sadness, and anxiety.

[0324] 5. Evaluations that reflect emotional states

[0325] The server will then evaluate the user's emotional state as determined by the emotion engine along with the detected fraud phrases. If the user is in a certain emotional state (e.g., anxiety or confusion), the likelihood of fraud increases, and points are added to the score.

[0326] 6. Alerting and Notification

[0327] Based on the results of the risk assessment, the server generates an alert, which includes a summary of the call, detected scam words, reputation score, text data, and the perceived emotional state. The alert is then sent to the user's contacts, including their next of kin and relevant authorities. Notifications can be sent via email, SMS, in-app notifications, and other methods.

[0328] 7. Always keep up with the latest criminal techniques

[0329] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[0330] Examples of prompts

[0331] For example, if a user includes a phrase such as "Please hurry up and transfer the money" during a call, while the emotion engine detects an emotional state such as anxiety or confusion, the smartphone device sends the voice data to the server in real time. The server converts the voice into text and detects the phrase "Please hurry up and transfer the money" and the user's sense of anxiety. The risk scoring algorithm evaluates these as high risk and immediately generates an alert to notify relatives and the police. This process makes it possible to prevent fraud before it occurs.

[0332] Example prompt sentence:

[0333] Create an application that immediately alerts relatives and the police if a user receives a phone call asking them to transfer money and feels uneasy. This application uses WebSocket to transmit voice data in real time and combines a speech recognition engine and an emotion recognition engine to evaluate fraudulent words and emotional state.

[0334] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0335] Step 1:

[0336] The terminal uses a microphone to collect call audio in real time from the moment the user answers the phone. The input is the user's speech, and the output is digitized audio data, which is sent to a server for further processing.

[0337] Step 2:

[0338] The device sends the collected audio data to the server using WebSocket. The input is the digitized audio data collected by the device, and the output is the audio data sent to the server in real time.

[0339] Step 3:

[0340] The server converts the received voice data into an appropriate format and inputs it to the voice recognition engine. The input is the voice data sent from the terminal, and the output is data in the appropriate format that is input to the voice recognition engine.

[0341] Step 4:

[0342] The server converts the voice data into text data using a speech recognition engine, where the input is the voice data converted into an appropriate format and the output is text data.

[0343] Step 5:

[0344] The server analyzes the text data generated by the speech recognition to detect specific fraudulent words and phrases. The input is the text data converted from speech, and the output is information containing the detected fraudulent words and phrases.

[0345] Step 6:

[0346] The server uses the detected fraud words and phrases to perform a risk assessment. The input is the detected fraud words and phrases, and the output is a fraud risk assessment score, which is used to generate subsequent alerts.

[0347] Step 7:

[0348] In parallel, the server inputs the voice data into the emotion engine to analyze the user's emotional state in real time. The input is the voice data received by the server, and the output is information identifying the user's emotional state (joy, anger, sadness, anxiety, etc.).

[0349] Step 8:

[0350] The server reflects the detected fraudulent words and the user's emotional state recognized by the emotion engine in the evaluation. The input is information on fraudulent words and emotional state, and the output is a detailed risk evaluation score that combines these.

[0351] Step 9:

[0352] Based on the risk assessment results, the server generates an alert, which includes a summary of the call, detected fraudulent words, the assessment score, text data, and the recognized emotional state. The input is the detailed risk assessment score and relevant data, and the output is the alert information.

[0353] Step 10:

[0354] The generated alerts are sent to pre-registered contacts of relatives and relevant authorities via email, SMS, in-app notifications, etc. The input is the generated alert information, and the output is the response of relatives and relevant authorities who receive the notification.

[0355] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0356] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0357] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0358] [Second embodiment]

[0359] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0360] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0361] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0362] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0363] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0364] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0365] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0366] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0367] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0368] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0369] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0370] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0371] This invention is a system for preventing special frauds. The system consists of terminal, server, and user components that work together to monitor and evaluate potentially fraudulent calls in real time, and alert relatives and relevant authorities as necessary.

[0372] Receiving and sending audio input

[0373] The device collects the call audio in real time from the moment the user answers the call. Audio input is captured using a microphone or similar audio input device, and the audio data is then transmitted to a server, typically using WebSocket or another real-time communication protocol.

[0374] Speech recognition and text conversion

[0375] The server receives the voice data sent from the device. The received voice data is converted into text data by a voice recognition engine. This voice recognition engine uses the latest machine learning technology, enabling highly accurate voice-to-text conversion.

[0376] Deceptive word detection and evaluation

[0377] The server analyzes the text data generated by the speech recognition to detect specific fraud words and phrases, typically using a predefined list of fraud-related keywords, which are then fed into a scoring algorithm to assess risk.

[0378] Alerting and Notifications

[0379] Based on the results of the risk assessment, the server generates an alert, which includes a summary of the call, detected scam words, and the text data of the call. The alert is then sent to pre-registered contacts, such as the user's relatives or relevant authorities. Notifications can be sent via email, SMS, in-app notifications, and other methods.

[0380] Always responding to the latest criminal techniques

[0381] The server continuously analyzes the received audio data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[0382] Specific examples

[0383] For example, if a user uses a phrase such as "Please hurry up and transfer the money" during a call, the device sends the voice data to the server in real time. The server converts the voice data into text and detects the phrase "Please hurry up and transfer the money." The risk scoring algorithm evaluates this phrase as high risk and immediately generates an alert to notify relatives and the police. This process can prevent fraud before it occurs.

[0384] This invention can effectively prevent damage caused by special frauds, and can significantly reduce the risk of being defrauded, especially for elderly people and single people. The specific programs and algorithms involved in implementing the invention will vary depending on the technology and environment to which it is applied, but they will generally be implemented based on the form shown here.

[0385] The processing flow will be explained below.

[0386] Step 1:

[0387] The terminal detects the call start event as soon as the user answers the phone and starts voice input. The voice is picked up in real time through a microphone and temporarily stored in a buffer as a digital signal.

[0388] Step 2:

[0389] The device transmits the captured audio data to the server in real time using WebSocket or other real-time communication protocols.

[0390] Step 3:

[0391] The server receives the voice data sent from the terminal, converts the received voice data into an appropriate format, and then inputs it into a voice recognition engine.

[0392] Step 4:

[0393] The server's speech recognition engine converts the speech data into text data. This engine performs highly accurate conversion from speech to text, and the generated text data is used for analysis.

[0394] Step 5:

[0395] The server analyzes the converted text data to detect specific fraudulent words and phrases using a pre-trained list of fraud-related keywords.

[0396] Step 6:

[0397] The server uses a scoring algorithm to assess the risk of detected scam words, with different weights assigned to different scam words to quantify their risk.

[0398] Step 7:

[0399] The server generates an alert based on the results of the scoring algorithm, which includes a summary of the call, detected scam words, reputation score, and text data.

[0400] Step 8:

[0401] The server then sends the generated alerts to pre-registered contacts of relatives and relevant authorities via email, SMS, in-app notifications, and more.

[0402] Step 9:

[0403] The server feeds the received voice and text data into a machine learning algorithm to constantly learn the latest fraud techniques, and the results of this learning are used for subsequent analyses.

[0404] Step 10:

[0405] The server stores all transaction logs in a database for future reference, including the start and end times of calls, analysis results, and notification content.

[0406] By going through the above steps, the present invention functions as an effective system for protecting users from special fraud.

[0407] Example 1

[0408] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0409] Currently, the methods of special fraud are evolving daily, increasing the risk of elderly people and single people falling victim to fraud. In particular, methods such as "bank transfer fraud" carried out over the phone are becoming more sophisticated, making it difficult for conventional prevention systems to deal with them. Therefore, a system is needed that analyzes the content of phone calls in real time, evaluates the risk of fraud, and issues alerts.

[0410] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0411] In this invention, the server includes a means for receiving voice input, a means for transmitting the received voice data to the server using a real-time communication protocol, and a means for converting the voice data into text data at the server, thereby enabling real-time analysis of the user's call content, detection of specific fraudulent words and phrases, and highly accurate risk assessment and prompt alert notification.

[0412] "Voice input" refers to a technology that uses a microphone or other voice input device to collect voices uttered by a user via a communication means such as a telephone.

[0413] A "real-time communication protocol" is a communication protocol for transmitting data such as telephone voice promptly with low latency. An example is WebSocket.

[0414] A "server" is a computer system that receives voice data sent from a terminal and performs processes such as voice recognition and data analysis.

[0415] "Text data" refers to character string information generated from voice data using voice recognition technology.

[0416] "Fraudulent words and phrases" are specific, predefined words or phrases related to fraudulent activity, such as "transfer" or "hurry."

[0417] A "scoring algorithm" is a calculation method used to numerically assess the risk of detected fraudulent words or phrases.

[0418] An "alert" is a warning message that is generated when dangerous call content is detected.

[0419] "Contacts" are people or organizations to whom alerts are sent. They are selected based on pre-registered information.

[0420] "Email" refers to digital letters sent and received over the Internet.

[0421] "SMS" is a means of sending and receiving short text messages over mobile phone networks.

[0422] "In-app notifications" refer to notification messages displayed through software such as smartphone apps.

[0423] The present invention is a system for preventing special frauds in advance, in which the components of the user, terminal, and server work together.

[0424] Receiving and sending audio input

[0425] The device collects the call audio in real time using a voice input device (e.g., a smartphone microphone) from the moment the user answers the call. The audio data is then sent to the server using a real-time communication protocol such as WebSocket.

[0426] Speech recognition and text conversion

[0427] The server receives the voice data sent from the device. The received voice data is converted into text data by a speech recognition engine (e.g., Google Cloud Speech-to-Text API). This speech recognition engine uses the latest machine learning technology, enabling highly accurate speech-to-text conversion.

[0428] Deceptive word detection and evaluation

[0429] The server analyzes the text data generated by the speech recognition. Natural Language Processing (NLP) algorithms are used for the analysis. The server uses a predefined list of fraud-related keywords to detect specific fraud words and phrases within the text data. The detected fraud words are then evaluated for risk using a scoring algorithm.

[0430] Alerting and Notifications

[0431] The server generates an alert based on the results of the risk assessment. The alert includes a summary of the call, detected scam words, and text data. The alert is then sent to pre-registered contacts, such as the user's relatives or relevant authorities, via email, SMS, or in-app notification.

[0432] Always responding to the latest criminal techniques

[0433] The server continuously analyzes the received audio data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[0434] Specific examples

[0435] For example, if a user uses a phrase such as "Please transfer the money quickly" during a call, the device will send the speech data to the server in real time. The server will convert the speech into text and detect the phrase "Please transfer the money quickly." The risk scoring algorithm will evaluate the phrase as high risk and immediately generate an alert to notify relatives and the police. This process can prevent fraud before it can be carried out.

[0436] Examples of prompt statements

[0437] "Detect specific phrases in the audio of a call and assess the likelihood of fraud. For example, if the phrase 'Please transfer the money quickly' is detected, issue a warning."

[0438] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0439] Step 1:

[0440] The device starts voice input.

[0441] The device uses a microphone to collect the call voice in real time from the moment the user answers the call. Specifically, the voice input software on the device runs in the background and collects the user's voice data.

[0442] Input: User's voice

[0443] Output: Collected audio data

[0444] Step 2:

[0445] Sending audio data to the server

[0446] The device sends the collected audio data to a server using a real-time communication protocol such as WebSocket, which requires stable communication with low latency.

[0447] Input: Collected audio data

[0448] Output: Audio data sent to the server

[0449] Step 3:

[0450] The server receives the audio data.

[0451] The server receives the voice data sent from the terminal in real time. The server process stores the received data and prepares it for subsequent voice recognition processing.

[0452] Input: Audio data sent from the device

[0453] Output: Archived audio data

[0454] Step 4:

[0455] Converting audio data to text

[0456] The server converts the voice data into text using a speech recognition engine, for example, the Google Cloud Speech-to-Text API, which ensures high accuracy in the process.

[0457] Input: Archived audio data

[0458] Output: Text data

[0459] Step 5:

[0460] Text data analysis

[0461] The server analyzes the text data generated by the speech recognition, using natural language processing (NLP) algorithms to detect fraudulent words and phrases, using a predefined list of fraudulent words.

[0462] Input: Text data

[0463] Output: Detected fraudulent words and phrases

[0464] Step 6:

[0465] Risk assessment

[0466] The server evaluates the detected fraudulent words and phrases using a scoring algorithm, which generates a numerical score indicating the likelihood of fraud.

[0467] Input: Detected scam words and phrases

[0468] Output: Risk score

[0469] Step 7:

[0470] Generate alerts

[0471] The server generates an alert based on the results of the risk assessment, which includes a summary of the call, detected scam words, and text data.

[0472] Input: Risk score, text data

[0473] Output: Alert

[0474] Step 8:

[0475] Alert Notification

[0476] Generated alerts are sent to pre-registered contacts via email, SMS, or in-app notifications.

[0477] Input: Alert

[0478] Output: Notified relatives and relevant authorities

[0479] Step 9:

[0480] Learning new fraud methods

[0481] The server uses the received voice data and analysis results to automatically learn new fraudulent techniques through machine learning algorithms, allowing the system to adapt to the latest fraud techniques.

[0482] Input: Received voice data, analysis results

[0483] Output: Updated deceptive wordlist and scoring algorithm

[0484] (Application example 1)

[0485] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0486] In modern society, special frauds are on the rise, with victims becoming increasingly serious, especially among the elderly. Telephone frauds are particularly sophisticated, leaving many victims vulnerable to deception. Current technology does not provide sufficient means to detect and prevent these frauds in real time. Therefore, a system is needed that analyzes the content of phone calls in real time and immediately sends an alert if there is a possibility of fraud. The system itself must also be able to learn and evolve so that it can constantly respond to the latest fraud techniques.

[0487] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0488] In this invention, the server includes a means for receiving voice input, a means for transmitting the received voice data to the server, a means for converting the voice data to text data at the server, a means for detecting specific phrases in the converted text data, a means for evaluating the risk level of the detected specific phrases, a means for generating an alert based on the evaluation results, a means for notifying pre-registered contacts of the generated alert, a means for collecting voice data and detecting fraudulent keywords in real time, and a means for automatically generating an alert to notify contacts when fraudulent keywords are detected. This allows the content of a user's calls to be analyzed in real time, and an alert to be immediately generated if there is a high possibility of fraud, making it possible to prevent damage before it occurs. Furthermore, by analyzing received voice data and constantly learning the latest criminal techniques, the system continuously evolves and can respond to more advanced fraud techniques.

[0489] "Audio input" refers to the means by which an audio signal is input to a digital device.

[0490] "Device" means an electronic device used by an individual or organization for communication and data processing.

[0491] "Server" refers to a computer system that receives, stores, analyzes, and transmits data over a network.

[0492] "Audio data" refers to data that represents an audio signal in digital form.

[0493] "Text data" refers to data obtained by converting voice data into character information.

[0494] "Specific phrases" refer to specific words or expressions that indicate possible fraud.

[0495] "Real-time" refers to data collection, analysis, notification, etc. being carried out immediately without delay.

[0496] "Risk assessment" refers to the process of analyzing and assessing the risk level of specific detected phrases.

[0497] An "alert" is a message that sends a warning or notification when a specific condition is met.

[0498] "Pre-registered Contacts" refers to contact information that has been pre-registered in the system for the purpose of sending alerts.

[0499] "Scam words" refer to specific keywords or phrases that indicate potential fraud.

[0500] "Machine learning" refers to technologies and methods that allow computers to analyze large amounts of data and learn automatically through algorithms.

[0501] The system of the present invention is designed to prevent special frauds, and involves the cooperation of the terminal, server, and user components. First, when a user starts a call, the terminal collects the user's voice input in real time and transmits it to the server as voice data. This voice data is transmitted using WebSocket or other real-time communication protocols.

[0502] The server converts the received voice data into text data using a speech recognition engine. This speech recognition engine uses the latest machine learning technology, such as Google's Speech-to-Text API, to achieve highly accurate speech-to-text conversion. The converted text data is then sent to the scam word detection and evaluation process.

[0503] To detect fraudulent words, a predefined list of fraud-related keywords is used. The server analyzes the generated text data to check whether it contains any of the keywords in this list. If any of these keywords are detected, a scoring algorithm is applied to assess the risk level. This algorithm evaluates the risk level of the detected keywords and generates an alert based on the results.

[0504] The generated alert will be sent to pre-registered contacts via email, SMS, or in-app notification, including a summary of the call and a recording. The alert will include information on the call that is likely to be fraudulent, allowing the user's relatives or relevant authorities to take immediate action. Notifications can be sent via notification services such as Amazon SNS.

[0505] Furthermore, the system uses machine learning algorithms to continuously analyze the received voice data in order to constantly adapt to the latest criminal techniques. This allows the system to respond quickly to new fraud methods. When a user actually experiences a call containing fraudulent words, the system accumulates that experience as learning data to improve future detection accuracy.

[0506] For example, if a user makes a call containing the phrase "Please hurry up and transfer the money," the device will send the voice data to the server in real time. The server will convert the voice data into text and detect the phrase "Please hurry up and transfer the money." Based on this detection result, a risk assessment will be performed and if the call is deemed high risk, an alert will be generated immediately and notified to relatives and relevant authorities. This process can prevent fraud from being carried out.

[0507] An example of a prompt sentence is, "There is a risk of important information being leaked during the phone call. Please evaluate the phrase 'Please transfer the money quickly' as a possible fraud and generate an alert."

[0508] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0509] Step 1:

[0510] When a user receives a call, the device collects the voice input through a microphone. This voice input is recorded as digital voice data. The input data is processed in real time and then sent to the next step. The output is the captured voice data.

[0511] Step 2:

[0512] The device sends the collected audio data to the server over the network. This process uses WebSocket or other real-time communication protocols. The input is the audio data collected on the device, and the output is the audio data sent to the server.

[0513] Step 3:

[0514] The server converts the received voice data into text data using a voice recognition engine, such as Google's Speech-to-Text API. The input is the voice data sent to the server, and the output is the converted text data.

[0515] Step 4:

[0516] The server analyzes the generated text data and detects fraudulent words using a predefined list of fraud-related keywords. It checks whether the text data contains any of the keywords in this list. The input is the text data, and the output is a list of detected fraudulent words.

[0517] Step 5:

[0518] The server applies a scoring algorithm to assess the risk of detected fraudulent words. The algorithm assesses the risk level of the detected keywords and outputs the result. The input is a list of fraudulent words, and the output is a risk assessment score.

[0519] Step 6:

[0520] Based on the assessment results, the server generates an alert, which is accompanied by information such as a summary of the call and the recording. The input is the risk assessment score and text data, and the output is the alert message.

[0521] Step 7:

[0522] Generated alerts are sent to pre-registered contacts via email, SMS, or in-app notification. This is done by the server using a notification protocol. The input is the alert message and the output is the notified alert.

[0523] Step 8:

[0524] To stay up to date with the latest criminal techniques, the server continuously analyzes the received audio data and uses machine learning algorithms to learn new fraud techniques. The input is previously processed audio data, and the output is an updated fraud wordlist and rating model.

[0525] This can prevent users from being involved in fraudulent activities and improve the accuracy of the system.

[0526] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0527] This invention combines an emotion engine with a system for preventing special frauds. The system's components, including the terminal, server, and user, work together to monitor and evaluate potentially fraudulent calls in real time, alerting relatives and relevant authorities as necessary. It also simultaneously recognizes the user's emotional state and incorporates that information into the evaluation, enabling more accurate fraud prevention.

[0528] Receiving and sending audio input

[0529] The device collects the call audio in real time from the moment the user answers the call. Audio input is captured using a microphone or similar audio input device, and the audio data is then transmitted to a server, typically using WebSocket or another real-time communication protocol.

[0530] Speech recognition and text conversion

[0531] The server receives the voice data sent from the device. The received voice data is first converted into an appropriate format and then input into a voice recognition engine. The voice recognition engine converts the voice data into text data. This engine performs highly accurate conversion from voice to text and the generated text data is used for analysis.

[0532] Deceptive word detection and evaluation

[0533] The server analyzes the text data generated by the speech recognition to detect specific fraud words and phrases using a predefined list of fraud-related keywords, which are then fed into a scoring algorithm to assess risk.

[0534] Emotion recognition by emotion engine

[0535] The server then inputs the received voice data into an emotion engine to analyze the user's emotional state in real time. The emotion engine analyzes changes in voice tone, speed, and volume to identify emotions such as joy, anger, sadness, and anxiety.

[0536] Evaluations that reflect emotional states

[0537] The server then evaluates the detected fraudulent phrases along with the user's emotional state as recognized by the emotion engine. If the user is in a certain emotional state (e.g., anxious or confused), points are added to the score to increase the likelihood of fraud.

[0538] Alerting and Notifications

[0539] Based on the results of the risk assessment, the server generates an alert, which includes a summary of the call, detected scam words, reputation score, text data, and the perceived emotional state. The alert is then sent to the user's contacts, including their next of kin and relevant authorities. Notifications can be sent via email, SMS, in-app notifications, and other methods.

[0540] Always responding to the latest criminal techniques

[0541] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[0542] Specific examples

[0543] For example, if a user uses a phrase like "Please hurry up and transfer the money" during a call, while the emotion engine detects an emotional state such as anxiety or confusion, the device sends the voice data to the server in real time. The server converts the voice into text and detects the phrase "Please hurry up and transfer the money" and the user's sense of anxiety. The risk scoring algorithm evaluates these as high risk and immediately generates an alert to notify relatives and the police. This process can prevent fraud from being committed.

[0544] This invention can effectively prevent damage caused by special frauds, and can significantly reduce the risk of being defrauded, especially for elderly people and single people. The specific programs and algorithms involved in implementing the invention will vary depending on the technology and environment to which it is applied, but they will generally be implemented based on the form shown here.

[0545] The processing flow will be explained below.

[0546] Step 1:

[0547] When the user answers the phone, the terminal detects the call start event and starts voice input. The call voice is picked up through the microphone and stored in a buffer in real time as a digital signal.

[0548] Step 2:

[0549] The device transmits the captured audio data to the server in real time, typically using WebSocket or other real-time communication protocols.

[0550] Step 3:

[0551] The server receives the voice data sent from the device and converts it into an appropriate format, which is then sent to a voice recognition engine.

[0552] Step 4:

[0553] The server's speech recognition engine converts the speech data into text data, which is then analyzed.

[0554] Step 5:

[0555] The server analyzes the generated text data to detect specific fraudulent words and phrases using a predefined list of fraud-related keywords.

[0556] Step 6:

[0557] The server uses a scoring algorithm to assess the risk of detected fraudulent words, which assigns different weights to different fraudulent words and quantifies their risk.

[0558] Step 7:

[0559] In parallel, the server inputs the voice data into an emotion engine to analyze the user's emotional state in real time, analyzing parameters such as voice tone, speed, and volume to identify emotions.

[0560] Step 8:

[0561] The server evaluates the risk level by taking into account the detected fraudulent phrases and the user's emotional state as recognized by the emotion engine. For example, if the user is anxious or confused, an evaluation score is added.

[0562] Step 9:

[0563] Based on the scoring algorithm, the server generates an alert that includes a summary of the call, detected scam words, reputation score, text data, and the perceived emotional state.

[0564] Step 10:

[0565] The generated alerts are sent by the server to pre-registered contacts via email, SMS, in-app notifications, and other methods.

[0566] Step 11:

[0567] The server feeds the received voice and text data into a machine learning algorithm to constantly learn the latest fraud techniques, allowing it to respond to new fraud methods.

[0568] Step 12:

[0569] The server stores all processing logs in a database, including the start and end times of calls, analysis results, and notification content.

[0570] By going through these steps, the system can protect users from special fraud and provide a fast and effective response.

[0571] Example 2

[0572] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0573] Special fraud methods are becoming more sophisticated every year, and the elderly and people living alone are at particularly high risk of falling victim to fraud. Conventional fraud prevention systems have the problem of being unable to completely prevent fraud victims because they have difficulty accurately identifying the content of fraud and are unable to reflect the user's emotional state in their evaluation.

[0574] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for receiving voice input, means for transmitting the received voice data to the server, means for converting the voice data into text data at the server, means for detecting specific phrases in the converted text data, means for evaluating the risk level of the detected specific phrases, means for analyzing the received voice data and recognizing the user's emotional state in real time, means for reflecting the user's emotional state in the evaluation, means for generating an alert based on the evaluation result, and means for notifying the generated alert to a pre-registered contact. This enables the possibility of fraud to be detected with high accuracy and enables a prompt and accurate response that also takes into account the user's emotional state.

[0575] "Means for receiving audio input" refers to a microphone or similar audio capture device for capturing audio data, such as speech.

[0576] "Means for transmitting received voice data to a server" refers to a method for transferring collected voice data to a server via a communication network such as the Internet.

[0577] "Means for converting voice data into text data at the server" refers to the process of converting voice data into text format using a voice recognition engine.

[0578] "Means for detecting specific phrases within the converted text data" refers to algorithms that search and detect pre-defined fraud-related keywords and phrases within the text data.

[0579] "Means for assessing the risk of specific detected phrases" refers to a method for quantifying and assessing the likelihood of fraud or the risk based on the detected phrases.

[0580] "Means for analyzing received voice data and recognizing the user's emotional state in real time" refers to an emotion engine that analyzes voice data and identifies the user's emotional state in real time, such as anxiety, anger, joy, or sadness.

[0581] "Means for reflecting the user's emotional state in the assessment" refers to a method for integrating and reflecting the user's emotional state recognized by the emotion engine in the fraud risk assessment.

[0582] "Means for generating alerts based on assessment results" refers to the process for generating warning messages or notifications when a high likelihood of fraud is assessed.

[0583] "Means of notifying pre-registered contacts of generated alerts" refers to means of sending generated alerts to pre-designated contacts, such as next of kin or relevant authorities, via email, SMS, in-app notifications, etc.

[0584] The present invention combines an emotion engine with a system for preventing special frauds, and is realized by the cooperation of the terminal, server, and user components.

[0585] Receiving and sending audio input

[0586] The device detects the start of a call when the user makes or receives a call. It catches the incoming call event and recognizes the beginning of the call. When the user starts a call, the device collects audio data of the conversation in real time using the built-in microphone or an external audio capture device. This audio data is sent to the server using a real-time communication technology such as the WebSocket protocol. To minimize communication delays, the audio data is sent intermittently, often using buffering.

[0587] Speech recognition and text conversion

[0588] The server receives the voice data sent from the device. The received voice data is stored in a specific buffer, where noise removal and voice normalization are performed as a preparatory step for analysis. The processed voice data is then input into a speech recognition engine (e.g., Google Speech-to-Text or IBM Watson) to convert the voice data into highly accurate text data. This text data is then analyzed.

[0589] Deceptive word detection and evaluation

[0590] The server analyzes the text data generated by the speech recognition engine. The analysis tool searches for and detects predefined fraud-related keywords and phrases (e.g., "transfer" or "hurry") within this text data. The detected fraud words are then input into a risk scoring algorithm, which evaluates the likelihood of the fraud.

[0591] Emotion recognition by emotion engine

[0592] In parallel, the server inputs the received voice data into the emotion engine to analyze the user's emotional state in real time. The emotion engine analyzes changes in voice tone, speed, and volume to identify the user's emotional state (e.g., anger, sadness, joy, anxiety, etc.).

[0593] Evaluations that reflect emotional states

[0594] The server integrates the detected fraud phrases with the user's emotional state as recognized by the emotion engine to evaluate the overall likelihood of fraud. In particular, if the user is anxious or confused, these feelings are added to the evaluation score.

[0595] Alerting and Notifications

[0596] Based on the evaluation results, the server generates an alert, which includes a summary of the call, detected scam words, evaluation score, text data, and the recognized emotional state, and the alert is sent to the user's relatives and relevant authorities via email, SMS, in-app notification, etc.

[0597] Always responding to the latest criminal techniques

[0598] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraud tactics, constantly updating the system's fraud detection capabilities.

[0599] Specific examples

[0600] For example, if a user says "Please hurry up and transfer the money" during a call, and the emotion engine detects the user's anxiety, the device will send the voice data to the server in real time. The server will convert the voice data into text, detect the phrase "Please hurry up and transfer the money," and the emotion engine will confirm the user's anxiety. The risk scoring algorithm will evaluate these as high risk and immediately generate an alert to notify relatives and the police. This process can prevent fraud from being committed.

[0601] Prompt Sentence Examples

[0602] Example prompts for input to a generative AI model:

[0603] "Please explain in detail the risk assessment process when the phrase 'hurry and transfer money' and the user's state of anxiety are detected during a call."

[0604] "Please tell me specifically how to send real-time collected call audio to a fraud detection system."

[0605] This system can significantly reduce the risk of elderly people and single people falling victim to fraud.

[0606] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0607] Step 1:

[0608] A call begins when a user makes or receives a call. The terminal catches this call reception event and recognizes the beginning of the call. The input includes the call reception event by the user. The output is the recognition of the start of the call.

[0609] Step 2:

[0610] When a user initiates a call, the device collects real-time audio data from the conversation using a built-in microphone or an external audio capture device. The input includes the user's voice during the call. The output is the collected audio data.

[0611] Step 3:

[0612] The device transmits the collected audio data to the server using real-time communication technology such as the WebSocket protocol. To minimize communication delays, the audio data is transmitted intermittently using buffering. The input includes the collected audio data. The output is the audio data transmitted to the server.

[0613] Step 4:

[0614] The server receives the audio data sent from the device. The received audio data is stored in a specific buffer, and noise removal and audio normalization are performed as a pre-analysis step. The input includes the audio data sent from the device. The output is the pre-processed audio data.

[0615] Step 5:

[0616] The server inputs the processed speech data into a speech recognition engine (e.g., Google Speech-to-Text or IBM Watson) to convert the speech data into text format. The input includes the preprocessed speech data. The output is the generated text data.

[0617] Step 6:

[0618] The server receives the generated text data from the speech recognition engine and inputs it into an analysis tool to detect predefined fraud-related keywords and phrases. The input includes the generated text data. The output is the detected fraud words and phrases.

[0619] Step 7:

[0620] The server inputs the detected fraud words into a risk scoring algorithm to quantify the likelihood of fraud. The input includes the detected fraud words or phrases. The output is a risk score.

[0621] Step 8:

[0622] In parallel, the server inputs the collected voice data into the emotion engine to analyze the user's emotional state in real time. The input includes the collected voice data. The output is the identification result of the user's emotional state.

[0623] Step 9:

[0624] The server integrates the detected scam phrases with the user's emotional state as recognized by the emotion engine to assess the overall likelihood of fraud. If the user is anxious or confused, this assessment is taken into account. Inputs include scam words, risk scores, and the user's emotional state. The output is the integrated assessment result.

[0625] Step 10:

[0626] Based on the evaluation results, the server generates an alert, which includes a summary of the call, detected scam words, the evaluation score, text data, and the recognized emotional state. The input includes the integrated evaluation results. The output is the generated alert.

[0627] Step 11:

[0628] The server notifies the generated alert to the user's relatives and relevant authorities via email, SMS, in-app notification, etc. The input includes the generated alert. The output is the notification sent.

[0629] Step 12:

[0630] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraud techniques, thereby constantly updating the system's fraud detection capabilities. Inputs include stored voice and text data. Outputs are the results of learning new criminal techniques.

[0631] (Application example 2)

[0632] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0633] Special frauds are becoming more sophisticated every year, and many people, especially the elderly, are falling victim to them. Frauds committed during phone calls have limited reaction time, making it difficult to respond immediately. Conventional fraud prevention systems detect fraud based on voice data, but do not take into account the user's psychological state, making it difficult to make accurate judgments. Since a rapid response to new fraud methods is also necessary, the present invention provides a system that solves these problems.

[0634] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving voice input, means for transmitting the received voice data to the server, means for converting the voice data to text data at the server, means for detecting specific phrases in the converted text data, means for evaluating the risk of the detected specific phrases, means for generating an alert based on the evaluation result, means for notifying pre-registered contacts of the generated alert, means for analyzing the user's emotional state, and means for reflecting the analyzed emotional state in a fraud risk evaluation. This enables highly accurate fraud detection and rapid response that takes the user's psychological state into consideration.

[0635] A "means for receiving voice input" is a device or system that converts a user's speech into an electrical signal and receives it in digital form.

[0636] The "means for transmitting received voice data to a server" refers to a method or protocol for transmitting voice data acquired by a terminal to a server via the Internet or other networks.

[0637] "Means for converting voice data into text data at the server" refers to the technology or algorithm that analyzes received voice data and converts the content into text format.

[0638] The "means for detecting specific phrases within the converted text data" refers to a filtering technique for identifying predefined specific words or phrases from the text data.

[0639] "Means for assessing the risk level of specific detected phrases" refers to algorithms or methods that assess the risk of the content of the call based on identified phrases and score them according to certain criteria.

[0640] The "means for generating alerts based on the evaluation results" is a system for creating warning messages and notifications based on the results of the risk evaluation.

[0641] "Means for notifying pre-registered contacts of generated alerts" refers to a mechanism for distributing generated warning messages to pre-defined contacts of next of kin and relevant authorities.

[0642] "Means for analyzing the user's emotional state" refers to technology that analyzes input data such as voice data and identifies the emotions (joy, anger, sadness, anxiety, etc.) felt by the user.

[0643] The "means for reflecting the analyzed emotional state in the fraud risk assessment" is a method for incorporating the user's emotional state into the risk assessment process and adjusting the risk score based on the results.

[0644] This invention combines an emotion engine with a system for preventing special frauds. Specifically, the system monitors users' calls in real time, assesses the possibility of fraud, and alerts relatives and relevant authorities if necessary.

[0645] The system is configured as follows:

[0646] Hardware and Software

[0647] The system is built using the user's smartphone, a server, and related software.

[0648] Hardware: Smartphone microphone, cellular network

[0649] Software: WebSocket (real-time transmission of voice data), requests (transmission of HTTP requests), speech-to-text API (conversion of voice data into text), emotion_recognition (analysis of emotional states), notification-service API (sending alerts)

[0650] Data processing and calculation

[0651] 1. Receiving and sending voice input

[0652] From the moment the user answers the call, the smartphone device collects the call audio in real time. Voice input is captured using a microphone, and then the audio data is sent to the server using WebSocket.

[0653] 2. Speech Recognition and Text Conversion

[0654] The server receives the voice data sent from the device. The received voice data is first converted into an appropriate format, and then converted into text data using a speech recognition engine. High-precision conversion from voice to text is performed, and the generated text data is then analyzed.

[0655] 3. Fraudulent word detection and evaluation

[0656] The server analyzes the text data generated by the speech recognition to detect specific fraud words and phrases using a predefined list of fraud-related keywords, which are then fed into a scoring algorithm to assess risk.

[0657] 4. Emotion Recognition by Emotion Engine

[0658] The server then inputs the received voice data into an emotion engine to analyze the user's emotional state in real time. The emotion engine analyzes changes in voice tone, speed, and volume to identify emotions such as joy, anger, sadness, and anxiety.

[0659] 5. Evaluations that reflect emotional states

[0660] The server will then evaluate the user's emotional state as determined by the emotion engine along with the detected fraud phrases. If the user is in a certain emotional state (e.g., anxiety or confusion), the likelihood of fraud increases, and points are added to the score.

[0661] 6. Alerting and Notification

[0662] Based on the results of the risk assessment, the server generates an alert, which includes a summary of the call, detected scam words, reputation score, text data, and the perceived emotional state. The alert is then sent to the user's contacts, including their next of kin and relevant authorities. Notifications can be sent via email, SMS, in-app notifications, and other methods.

[0663] 7. Always keep up with the latest criminal techniques

[0664] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[0665] Examples of prompts

[0666] For example, if a user includes a phrase such as "Please hurry up and transfer the money" during a call, while the emotion engine detects an emotional state such as anxiety or confusion, the smartphone device sends the voice data to the server in real time. The server converts the voice into text and detects the phrase "Please hurry up and transfer the money" and the user's sense of anxiety. The risk scoring algorithm evaluates these as high risk and immediately generates an alert to notify relatives and the police. This process makes it possible to prevent fraud before it occurs.

[0667] Example prompt sentence:

[0668] Create an application that immediately alerts relatives and the police if a user receives a phone call asking them to transfer money and feels uneasy. This application uses WebSocket to transmit voice data in real time and combines a speech recognition engine and an emotion recognition engine to evaluate fraudulent words and emotional state.

[0669] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0670] Step 1:

[0671] The terminal uses a microphone to collect call audio in real time from the moment the user answers the phone. The input is the user's speech, and the output is digitized audio data, which is sent to a server for further processing.

[0672] Step 2:

[0673] The device sends the collected audio data to the server using WebSocket. The input is the digitized audio data collected by the device, and the output is the audio data sent to the server in real time.

[0674] Step 3:

[0675] The server converts the received voice data into an appropriate format and inputs it to the voice recognition engine. The input is the voice data sent from the terminal, and the output is data in the appropriate format that is input to the voice recognition engine.

[0676] Step 4:

[0677] The server converts the voice data into text data using a speech recognition engine, where the input is the voice data converted into an appropriate format and the output is text data.

[0678] Step 5:

[0679] The server analyzes the text data generated by the speech recognition to detect specific fraudulent words and phrases. The input is the text data converted from speech, and the output is information containing the detected fraudulent words and phrases.

[0680] Step 6:

[0681] The server uses the detected fraud words and phrases to perform a risk assessment. The input is the detected fraud words and phrases, and the output is a fraud risk assessment score, which is used to generate subsequent alerts.

[0682] Step 7:

[0683] In parallel, the server inputs the voice data into the emotion engine to analyze the user's emotional state in real time. The input is the voice data received by the server, and the output is information identifying the user's emotional state (joy, anger, sadness, anxiety, etc.).

[0684] Step 8:

[0685] The server reflects the detected fraudulent words and the user's emotional state recognized by the emotion engine in the evaluation. The input is information on fraudulent words and emotional state, and the output is a detailed risk evaluation score that combines these.

[0686] Step 9:

[0687] Based on the risk assessment results, the server generates an alert, which includes a summary of the call, detected fraudulent words, the assessment score, text data, and the recognized emotional state. The input is the detailed risk assessment score and relevant data, and the output is the alert information.

[0688] Step 10:

[0689] The generated alerts are sent to pre-registered contacts of relatives and relevant authorities via email, SMS, in-app notifications, etc. The input is the generated alert information, and the output is the response of relatives and relevant authorities who receive the notification.

[0690] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0691] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0692] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0693] [Third embodiment]

[0694] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0695] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0696] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0697] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0698] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0699] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0700] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0701] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0702] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0703] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0704] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0705] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0706] This invention is a system for preventing special frauds. The system consists of terminal, server, and user components that work together to monitor and evaluate potentially fraudulent calls in real time, and alert relatives and relevant authorities as necessary.

[0707] Receiving and sending audio input

[0708] The device collects the call audio in real time from the moment the user answers the call. Audio input is captured using a microphone or similar audio input device, and the audio data is then transmitted to a server, typically using WebSocket or another real-time communication protocol.

[0709] Speech recognition and text conversion

[0710] The server receives the voice data sent from the device. The received voice data is converted into text data by a voice recognition engine. This voice recognition engine uses the latest machine learning technology, enabling highly accurate voice-to-text conversion.

[0711] Deceptive word detection and evaluation

[0712] The server analyzes the text data generated by the speech recognition to detect specific fraud words and phrases, typically using a predefined list of fraud-related keywords, which are then fed into a scoring algorithm to assess risk.

[0713] Alerting and Notifications

[0714] Based on the results of the risk assessment, the server generates an alert, which includes a summary of the call, detected scam words, and the text data of the call. The alert is then sent to pre-registered contacts, such as the user's relatives or relevant authorities. Notifications can be sent via email, SMS, in-app notifications, and other methods.

[0715] Always responding to the latest criminal techniques

[0716] The server continuously analyzes the received audio data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[0717] Specific examples

[0718] For example, if a user uses a phrase such as "Please hurry up and transfer the money" during a call, the device sends the voice data to the server in real time. The server converts the voice data into text and detects the phrase "Please hurry up and transfer the money." The risk scoring algorithm evaluates this phrase as high risk and immediately generates an alert to notify relatives and the police. This process can prevent fraud before it occurs.

[0719] This invention can effectively prevent damage caused by special frauds, and can significantly reduce the risk of being defrauded, especially for elderly people and single people. The specific programs and algorithms involved in implementing the invention will vary depending on the technology and environment to which it is applied, but they will generally be implemented based on the form shown here.

[0720] The processing flow will be explained below.

[0721] Step 1:

[0722] The terminal detects the call start event as soon as the user answers the phone and starts voice input. The voice is picked up in real time through a microphone and temporarily stored in a buffer as a digital signal.

[0723] Step 2:

[0724] The device transmits the captured audio data to the server in real time using WebSocket or other real-time communication protocols.

[0725] Step 3:

[0726] The server receives the voice data sent from the terminal, converts the received voice data into an appropriate format, and then inputs it into a voice recognition engine.

[0727] Step 4:

[0728] The server's speech recognition engine converts the speech data into text data. This engine performs highly accurate conversion from speech to text, and the generated text data is used for analysis.

[0729] Step 5:

[0730] The server analyzes the converted text data to detect specific fraudulent words and phrases using a pre-trained list of fraud-related keywords.

[0731] Step 6:

[0732] The server uses a scoring algorithm to assess the risk of detected scam words, with different weights assigned to different scam words to quantify their risk.

[0733] Step 7:

[0734] The server generates an alert based on the results of the scoring algorithm, which includes a summary of the call, detected scam words, reputation score, and text data.

[0735] Step 8:

[0736] The server then sends the generated alerts to pre-registered contacts of relatives and relevant authorities via email, SMS, in-app notifications, and more.

[0737] Step 9:

[0738] The server feeds the received voice and text data into a machine learning algorithm to constantly learn the latest fraud techniques, and the results of this learning are used for subsequent analyses.

[0739] Step 10:

[0740] The server stores all transaction logs in a database for future reference, including the start and end times of calls, analysis results, and notification content.

[0741] By going through the above steps, the present invention functions as an effective system for protecting users from special fraud.

[0742] Example 1

[0743] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0744] Currently, the methods of special fraud are evolving daily, increasing the risk of elderly people and single people falling victim to fraud. In particular, methods such as "bank transfer fraud" carried out over the phone are becoming more sophisticated, making it difficult for conventional prevention systems to deal with them. Therefore, a system is needed that analyzes the content of phone calls in real time, evaluates the risk of fraud, and issues alerts.

[0745] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0746] In this invention, the server includes a means for receiving voice input, a means for transmitting the received voice data to the server using a real-time communication protocol, and a means for converting the voice data into text data at the server, thereby enabling real-time analysis of the user's call content, detection of specific fraudulent words and phrases, and highly accurate risk assessment and prompt alert notification.

[0747] "Voice input" refers to a technology that uses a microphone or other voice input device to collect voices uttered by a user via a communication means such as a telephone.

[0748] A "real-time communication protocol" is a communication protocol for transmitting data such as telephone voice promptly with low latency. An example is WebSocket.

[0749] A "server" is a computer system that receives voice data sent from a terminal and performs processes such as voice recognition and data analysis.

[0750] "Text data" refers to character string information generated from voice data using voice recognition technology.

[0751] "Fraudulent words and phrases" are specific, predefined words or phrases related to fraudulent activity, such as "transfer" or "hurry."

[0752] A "scoring algorithm" is a calculation method used to numerically assess the risk of detected fraudulent words or phrases.

[0753] An "alert" is a warning message that is generated when dangerous call content is detected.

[0754] "Contacts" are people or organizations to whom alerts are sent. They are selected based on pre-registered information.

[0755] "Email" refers to digital letters sent and received over the Internet.

[0756] "SMS" is a means of sending and receiving short text messages over mobile phone networks.

[0757] "In-app notifications" refer to notification messages displayed through software such as smartphone apps.

[0758] The present invention is a system for preventing special frauds in advance, in which the components of the user, terminal, and server work together.

[0759] Receiving and sending audio input

[0760] The device collects the call audio in real time using a voice input device (e.g., a smartphone microphone) from the moment the user answers the call. The audio data is then sent to the server using a real-time communication protocol such as WebSocket.

[0761] Speech recognition and text conversion

[0762] The server receives the voice data sent from the device. The received voice data is converted into text data by a speech recognition engine (e.g., Google Cloud Speech-to-Text API). This speech recognition engine uses the latest machine learning technology, enabling highly accurate speech-to-text conversion.

[0763] Deceptive word detection and evaluation

[0764] The server analyzes the text data generated by the speech recognition. Natural Language Processing (NLP) algorithms are used for the analysis. The server uses a predefined list of fraud-related keywords to detect specific fraud words and phrases within the text data. The detected fraud words are then evaluated for risk using a scoring algorithm.

[0765] Alerting and Notifications

[0766] The server generates an alert based on the results of the risk assessment. The alert includes a summary of the call, detected scam words, and text data. The alert is then sent to pre-registered contacts, such as the user's relatives or relevant authorities, via email, SMS, or in-app notification.

[0767] Always responding to the latest criminal techniques

[0768] The server continuously analyzes the received audio data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[0769] Specific examples

[0770] For example, if a user uses a phrase such as "Please transfer the money quickly" during a call, the device will send the speech data to the server in real time. The server will convert the speech into text and detect the phrase "Please transfer the money quickly." The risk scoring algorithm will evaluate the phrase as high risk and immediately generate an alert to notify relatives and the police. This process can prevent fraud before it can be carried out.

[0771] Examples of prompt statements

[0772] "Detect specific phrases in the audio of a call and assess the likelihood of fraud. For example, if the phrase 'Please transfer the money quickly' is detected, issue a warning."

[0773] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0774] Step 1:

[0775] The device starts voice input.

[0776] The device uses a microphone to collect the call voice in real time from the moment the user answers the call. Specifically, the voice input software on the device runs in the background and collects the user's voice data.

[0777] Input: User's voice

[0778] Output: Collected audio data

[0779] Step 2:

[0780] Sending audio data to the server

[0781] The device sends the collected audio data to a server using a real-time communication protocol such as WebSocket, which requires stable communication with low latency.

[0782] Input: Collected audio data

[0783] Output: Audio data sent to the server

[0784] Step 3:

[0785] The server receives the audio data.

[0786] The server receives the voice data sent from the terminal in real time. The server process stores the received data and prepares it for subsequent voice recognition processing.

[0787] Input: Audio data sent from the device

[0788] Output: Archived audio data

[0789] Step 4:

[0790] Converting audio data to text

[0791] The server converts the voice data into text using a speech recognition engine, for example, the Google Cloud Speech-to-Text API, which ensures high accuracy in the process.

[0792] Input: Archived audio data

[0793] Output: Text data

[0794] Step 5:

[0795] Text data analysis

[0796] The server analyzes the text data generated by the speech recognition, using natural language processing (NLP) algorithms to detect fraudulent words and phrases, using a predefined list of fraudulent words.

[0797] Input: Text data

[0798] Output: Detected fraudulent words and phrases

[0799] Step 6:

[0800] Risk assessment

[0801] The server evaluates the detected fraudulent words and phrases using a scoring algorithm, which generates a numerical score indicating the likelihood of fraud.

[0802] Input: Detected scam words and phrases

[0803] Output: Risk score

[0804] Step 7:

[0805] Generate alerts

[0806] The server generates an alert based on the results of the risk assessment, which includes a summary of the call, detected scam words, and text data.

[0807] Input: Risk score, text data

[0808] Output: Alert

[0809] Step 8:

[0810] Alert Notification

[0811] Generated alerts are sent to pre-registered contacts via email, SMS, or in-app notifications.

[0812] Input: Alert

[0813] Output: Notified relatives and relevant authorities

[0814] Step 9:

[0815] Learning new fraud methods

[0816] The server uses the received voice data and analysis results to automatically learn new fraudulent techniques through machine learning algorithms, allowing the system to adapt to the latest fraud techniques.

[0817] Input: Received voice data, analysis results

[0818] Output: Updated deceptive wordlist and scoring algorithm

[0819] (Application example 1)

[0820] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0821] In modern society, special frauds are on the rise, with victims becoming increasingly serious, especially among the elderly. Telephone frauds are particularly sophisticated, leaving many victims vulnerable to deception. Current technology does not provide sufficient means to detect and prevent these frauds in real time. Therefore, a system is needed that analyzes the content of phone calls in real time and immediately sends an alert if there is a possibility of fraud. The system itself must also be able to learn and evolve so that it can constantly respond to the latest fraud techniques.

[0822] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0823] In this invention, the server includes a means for receiving voice input, a means for transmitting the received voice data to the server, a means for converting the voice data to text data at the server, a means for detecting specific phrases in the converted text data, a means for evaluating the risk level of the detected specific phrases, a means for generating an alert based on the evaluation results, a means for notifying pre-registered contacts of the generated alert, a means for collecting voice data and detecting fraudulent keywords in real time, and a means for automatically generating an alert to notify contacts when fraudulent keywords are detected. This allows the content of a user's calls to be analyzed in real time, and an alert to be immediately generated if there is a high possibility of fraud, making it possible to prevent damage before it occurs. Furthermore, by analyzing received voice data and constantly learning the latest criminal techniques, the system continuously evolves and can respond to more advanced fraud techniques.

[0824] "Audio input" refers to the means by which an audio signal is input to a digital device.

[0825] "Device" means an electronic device used by an individual or organization for communication and data processing.

[0826] "Server" refers to a computer system that receives, stores, analyzes, and transmits data over a network.

[0827] "Audio data" refers to data that represents an audio signal in digital form.

[0828] "Text data" refers to data obtained by converting voice data into character information.

[0829] "Specific phrases" refer to specific words or expressions that indicate possible fraud.

[0830] "Real-time" refers to data collection, analysis, notification, etc. being carried out immediately without delay.

[0831] "Risk assessment" refers to the process of analyzing and assessing the risk level of specific detected phrases.

[0832] An "alert" is a message that sends a warning or notification when a specific condition is met.

[0833] "Pre-registered Contacts" refers to contact information that has been pre-registered in the system for the purpose of sending alerts.

[0834] "Scam words" refer to specific keywords or phrases that indicate potential fraud.

[0835] "Machine learning" refers to technologies and methods that allow computers to analyze large amounts of data and learn automatically through algorithms.

[0836] The system of the present invention is designed to prevent special frauds, and involves the cooperation of the terminal, server, and user components. First, when a user starts a call, the terminal collects the user's voice input in real time and transmits it to the server as voice data. This voice data is transmitted using WebSocket or other real-time communication protocols.

[0837] The server converts the received voice data into text data using a speech recognition engine. This speech recognition engine uses the latest machine learning technology, such as Google's Speech-to-Text API, to achieve highly accurate speech-to-text conversion. The converted text data is then sent to the scam word detection and evaluation process.

[0838] To detect fraudulent words, a predefined list of fraud-related keywords is used. The server analyzes the generated text data to check whether it contains any of the keywords in this list. If any of these keywords are detected, a scoring algorithm is applied to assess the risk level. This algorithm evaluates the risk level of the detected keywords and generates an alert based on the results.

[0839] The generated alert will be sent to pre-registered contacts via email, SMS, or in-app notification, including a summary of the call and a recording. The alert will include information on the call that is likely to be fraudulent, allowing the user's relatives or relevant authorities to take immediate action. Notifications can be sent via notification services such as Amazon SNS.

[0840] Furthermore, the system uses machine learning algorithms to continuously analyze the received voice data in order to constantly adapt to the latest criminal techniques. This allows the system to respond quickly to new fraud methods. When a user actually experiences a call containing fraudulent words, the system accumulates that experience as learning data to improve future detection accuracy.

[0841] For example, if a user makes a call containing the phrase "Please hurry up and transfer the money," the device will send the voice data to the server in real time. The server will convert the voice data into text and detect the phrase "Please hurry up and transfer the money." Based on this detection result, a risk assessment will be performed and if the call is deemed high risk, an alert will be generated immediately and notified to relatives and relevant authorities. This process can prevent fraud from being carried out.

[0842] An example of a prompt sentence is, "There is a risk of important information being leaked during the phone call. Please evaluate the phrase 'Please transfer the money quickly' as a possible fraud and generate an alert."

[0843] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0844] Step 1:

[0845] When a user receives a call, the device collects the voice input through a microphone. This voice input is recorded as digital voice data. The input data is processed in real time and then sent to the next step. The output is the captured voice data.

[0846] Step 2:

[0847] The device sends the collected audio data to the server over the network. This process uses WebSocket or other real-time communication protocols. The input is the audio data collected on the device, and the output is the audio data sent to the server.

[0848] Step 3:

[0849] The server converts the received voice data into text data using a voice recognition engine, such as Google's Speech-to-Text API. The input is the voice data sent to the server, and the output is the converted text data.

[0850] Step 4:

[0851] The server analyzes the generated text data and detects fraudulent words using a predefined list of fraud-related keywords. It checks whether the text data contains any of the keywords in this list. The input is the text data, and the output is a list of detected fraudulent words.

[0852] Step 5:

[0853] The server applies a scoring algorithm to assess the risk of detected fraudulent words. The algorithm assesses the risk level of the detected keywords and outputs the result. The input is a list of fraudulent words, and the output is a risk assessment score.

[0854] Step 6:

[0855] Based on the assessment results, the server generates an alert, which is accompanied by information such as a summary of the call and the recording. The input is the risk assessment score and text data, and the output is the alert message.

[0856] Step 7:

[0857] Generated alerts are sent to pre-registered contacts via email, SMS, or in-app notification. This is done by the server using a notification protocol. The input is the alert message and the output is the notified alert.

[0858] Step 8:

[0859] To stay up to date with the latest criminal techniques, the server continuously analyzes the received audio data and uses machine learning algorithms to learn new fraud techniques. The input is previously processed audio data, and the output is an updated fraud wordlist and rating model.

[0860] This can prevent users from being involved in fraudulent activities and improve the accuracy of the system.

[0861] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0862] This invention combines an emotion engine with a system for preventing special frauds. The system's components, including the terminal, server, and user, work together to monitor and evaluate potentially fraudulent calls in real time, alerting relatives and relevant authorities as necessary. It also simultaneously recognizes the user's emotional state and incorporates that information into the evaluation, enabling more accurate fraud prevention.

[0863] Receiving and sending audio input

[0864] The device collects the call audio in real time from the moment the user answers the call. Audio input is captured using a microphone or similar audio input device, and the audio data is then transmitted to a server, typically using WebSocket or another real-time communication protocol.

[0865] Speech recognition and text conversion

[0866] The server receives the voice data sent from the device. The received voice data is first converted into an appropriate format and then input into a voice recognition engine. The voice recognition engine converts the voice data into text data. This engine performs highly accurate conversion from voice to text and the generated text data is used for analysis.

[0867] Deceptive word detection and evaluation

[0868] The server analyzes the text data generated by the speech recognition to detect specific fraud words and phrases using a predefined list of fraud-related keywords, which are then fed into a scoring algorithm to assess risk.

[0869] Emotion recognition by emotion engine

[0870] The server then inputs the received voice data into an emotion engine to analyze the user's emotional state in real time. The emotion engine analyzes changes in voice tone, speed, and volume to identify emotions such as joy, anger, sadness, and anxiety.

[0871] Evaluations that reflect emotional states

[0872] The server then evaluates the detected fraudulent phrases along with the user's emotional state as recognized by the emotion engine. If the user is in a certain emotional state (e.g., anxious or confused), points are added to the score to increase the likelihood of fraud.

[0873] Alerting and Notifications

[0874] Based on the results of the risk assessment, the server generates an alert, which includes a summary of the call, detected scam words, reputation score, text data, and the perceived emotional state. The alert is then sent to the user's contacts, including their next of kin and relevant authorities. Notifications can be sent via email, SMS, in-app notifications, and other methods.

[0875] Always responding to the latest criminal techniques

[0876] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[0877] Specific examples

[0878] For example, if a user uses a phrase like "Please hurry up and transfer the money" during a call, while the emotion engine detects an emotional state such as anxiety or confusion, the device sends the voice data to the server in real time. The server converts the voice into text and detects the phrase "Please hurry up and transfer the money" and the user's sense of anxiety. The risk scoring algorithm evaluates these as high risk and immediately generates an alert to notify relatives and the police. This process can prevent fraud from being committed.

[0879] This invention can effectively prevent damage caused by special frauds, and can significantly reduce the risk of being defrauded, especially for elderly people and single people. The specific programs and algorithms involved in implementing the invention will vary depending on the technology and environment to which it is applied, but they will generally be implemented based on the form shown here.

[0880] The processing flow will be explained below.

[0881] Step 1:

[0882] When the user answers the phone, the terminal detects the call start event and starts voice input. The call voice is picked up through the microphone and stored in a buffer in real time as a digital signal.

[0883] Step 2:

[0884] The device transmits the captured audio data to the server in real time, typically using WebSocket or other real-time communication protocols.

[0885] Step 3:

[0886] The server receives the voice data sent from the device and converts it into an appropriate format, which is then sent to a voice recognition engine.

[0887] Step 4:

[0888] The server's speech recognition engine converts the speech data into text data, which is then analyzed.

[0889] Step 5:

[0890] The server analyzes the generated text data to detect specific fraudulent words and phrases using a predefined list of fraud-related keywords.

[0891] Step 6:

[0892] The server uses a scoring algorithm to assess the risk of detected fraudulent words, which assigns different weights to different fraudulent words and quantifies their risk.

[0893] Step 7:

[0894] In parallel, the server inputs the voice data into an emotion engine to analyze the user's emotional state in real time, analyzing parameters such as voice tone, speed, and volume to identify emotions.

[0895] Step 8:

[0896] The server evaluates the risk level by taking into account the detected fraudulent phrases and the user's emotional state as recognized by the emotion engine. For example, if the user is anxious or confused, an evaluation score is added.

[0897] Step 9:

[0898] Based on the scoring algorithm, the server generates an alert that includes a summary of the call, detected scam words, reputation score, text data, and the perceived emotional state.

[0899] Step 10:

[0900] The generated alerts are sent by the server to pre-registered contacts via email, SMS, in-app notifications, and other methods.

[0901] Step 11:

[0902] The server feeds the received voice and text data into a machine learning algorithm to constantly learn the latest fraud techniques, allowing it to respond to new fraud methods.

[0903] Step 12:

[0904] The server stores all processing logs in a database, including the start and end times of calls, analysis results, and notification content.

[0905] By going through these steps, the system can protect users from special fraud and provide a fast and effective response.

[0906] Example 2

[0907] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0908] Special fraud methods are becoming more sophisticated every year, and the elderly and people living alone are at particularly high risk of falling victim to fraud. Conventional fraud prevention systems have the problem of being unable to completely prevent fraud victims because they have difficulty accurately identifying the content of fraud and are unable to reflect the user's emotional state in their evaluation.

[0909] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for receiving voice input, means for transmitting the received voice data to the server, means for converting the voice data into text data at the server, means for detecting specific phrases in the converted text data, means for evaluating the risk level of the detected specific phrases, means for analyzing the received voice data and recognizing the user's emotional state in real time, means for reflecting the user's emotional state in the evaluation, means for generating an alert based on the evaluation result, and means for notifying the generated alert to a pre-registered contact. This enables the possibility of fraud to be detected with high accuracy and enables a prompt and accurate response that also takes into account the user's emotional state.

[0910] "Means for receiving audio input" refers to a microphone or similar audio capture device for capturing audio data, such as speech.

[0911] "Means for transmitting received voice data to a server" refers to a method for transferring collected voice data to a server via a communication network such as the Internet.

[0912] "Means for converting voice data into text data at the server" refers to the process of converting voice data into text format using a voice recognition engine.

[0913] "Means for detecting specific phrases within the converted text data" refers to algorithms that search and detect pre-defined fraud-related keywords and phrases within the text data.

[0914] "Means for assessing the risk of specific detected phrases" refers to a method for quantifying and assessing the likelihood of fraud or the risk based on the detected phrases.

[0915] "Means for analyzing received voice data and recognizing the user's emotional state in real time" refers to an emotion engine that analyzes voice data and identifies the user's emotional state in real time, such as anxiety, anger, joy, or sadness.

[0916] "Means for reflecting the user's emotional state in the assessment" refers to a method for integrating and reflecting the user's emotional state recognized by the emotion engine in the fraud risk assessment.

[0917] "Means for generating alerts based on assessment results" refers to the process for generating warning messages or notifications when a high likelihood of fraud is assessed.

[0918] "Means of notifying pre-registered contacts of generated alerts" refers to means of sending generated alerts to pre-designated contacts, such as next of kin or relevant authorities, via email, SMS, in-app notifications, etc.

[0919] The present invention combines an emotion engine with a system for preventing special frauds, and is realized by the cooperation of the terminal, server, and user components.

[0920] Receiving and sending audio input

[0921] The device detects the start of a call when the user makes or receives a call. It catches the incoming call event and recognizes the beginning of the call. When the user starts a call, the device collects audio data of the conversation in real time using the built-in microphone or an external audio capture device. This audio data is sent to the server using a real-time communication technology such as the WebSocket protocol. To minimize communication delays, the audio data is sent intermittently, often using buffering.

[0922] Speech recognition and text conversion

[0923] The server receives the voice data sent from the device. The received voice data is stored in a specific buffer, where noise removal and voice normalization are performed as a preparatory step for analysis. The processed voice data is then input into a speech recognition engine (e.g., Google Speech-to-Text or IBM Watson) to convert the voice data into highly accurate text data. This text data is then analyzed.

[0924] Deceptive word detection and evaluation

[0925] The server analyzes the text data generated by the speech recognition engine. The analysis tool searches for and detects predefined fraud-related keywords and phrases (e.g., "transfer" or "hurry") within this text data. The detected fraud words are then input into a risk scoring algorithm, which evaluates the likelihood of the fraud.

[0926] Emotion recognition by emotion engine

[0927] In parallel, the server inputs the received voice data into the emotion engine to analyze the user's emotional state in real time. The emotion engine analyzes changes in voice tone, speed, and volume to identify the user's emotional state (e.g., anger, sadness, joy, anxiety, etc.).

[0928] Evaluations that reflect emotional states

[0929] The server integrates the detected fraud phrases with the user's emotional state as recognized by the emotion engine to evaluate the overall likelihood of fraud. In particular, if the user is anxious or confused, these feelings are added to the evaluation score.

[0930] Alerting and Notifications

[0931] Based on the evaluation results, the server generates an alert, which includes a summary of the call, detected scam words, evaluation score, text data, and the recognized emotional state, and the alert is sent to the user's relatives and relevant authorities via email, SMS, in-app notification, etc.

[0932] Always responding to the latest criminal techniques

[0933] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraud tactics, constantly updating the system's fraud detection capabilities.

[0934] Specific examples

[0935] For example, if a user says "Please hurry up and transfer the money" during a call, and the emotion engine detects the user's anxiety, the device will send the voice data to the server in real time. The server will convert the voice data into text, detect the phrase "Please hurry up and transfer the money," and the emotion engine will confirm the user's anxiety. The risk scoring algorithm will evaluate these as high risk and immediately generate an alert to notify relatives and the police. This process can prevent fraud from being committed.

[0936] Prompt Sentence Examples

[0937] Example prompts for input to a generative AI model:

[0938] "Please explain in detail the risk assessment process when the phrase 'hurry and transfer money' and the user's state of anxiety are detected during a call."

[0939] "Please tell me specifically how to send real-time collected call audio to a fraud detection system."

[0940] This system can significantly reduce the risk of elderly people and single people falling victim to fraud.

[0941] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0942] Step 1:

[0943] A call begins when a user makes or receives a call. The terminal catches this call reception event and recognizes the beginning of the call. The input includes the call reception event by the user. The output is the recognition of the start of the call.

[0944] Step 2:

[0945] When a user initiates a call, the device collects real-time audio data from the conversation using a built-in microphone or an external audio capture device. The input includes the user's voice during the call. The output is the collected audio data.

[0946] Step 3:

[0947] The device transmits the collected audio data to the server using real-time communication technology such as the WebSocket protocol. To minimize communication delays, the audio data is transmitted intermittently using buffering. The input includes the collected audio data. The output is the audio data transmitted to the server.

[0948] Step 4:

[0949] The server receives the audio data sent from the device. The received audio data is stored in a specific buffer, and noise removal and audio normalization are performed as a pre-analysis step. The input includes the audio data sent from the device. The output is the pre-processed audio data.

[0950] Step 5:

[0951] The server inputs the processed speech data into a speech recognition engine (e.g., Google Speech-to-Text or IBM Watson) to convert the speech data into text format. The input includes the preprocessed speech data. The output is the generated text data.

[0952] Step 6:

[0953] The server receives the generated text data from the speech recognition engine and inputs it into an analysis tool to detect predefined fraud-related keywords and phrases. The input includes the generated text data. The output is the detected fraud words and phrases.

[0954] Step 7:

[0955] The server inputs the detected fraud words into a risk scoring algorithm to quantify the likelihood of fraud. The input includes the detected fraud words or phrases. The output is a risk score.

[0956] Step 8:

[0957] In parallel, the server inputs the collected voice data into the emotion engine to analyze the user's emotional state in real time. The input includes the collected voice data. The output is the identification result of the user's emotional state.

[0958] Step 9:

[0959] The server integrates the detected scam phrases with the user's emotional state as recognized by the emotion engine to assess the overall likelihood of fraud. If the user is anxious or confused, this assessment is taken into account. Inputs include scam words, risk scores, and the user's emotional state. The output is the integrated assessment result.

[0960] Step 10:

[0961] Based on the evaluation results, the server generates an alert, which includes a summary of the call, detected scam words, the evaluation score, text data, and the recognized emotional state. The input includes the integrated evaluation results. The output is the generated alert.

[0962] Step 11:

[0963] The server notifies the generated alert to the user's relatives and relevant authorities via email, SMS, in-app notification, etc. The input includes the generated alert. The output is the notification sent.

[0964] Step 12:

[0965] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraud techniques, thereby constantly updating the system's fraud detection capabilities. Inputs include stored voice and text data. Outputs are the results of learning new criminal techniques.

[0966] (Application example 2)

[0967] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0968] Special frauds are becoming more sophisticated every year, and many people, especially the elderly, are falling victim to them. Frauds committed during phone calls have limited reaction time, making it difficult to respond immediately. Conventional fraud prevention systems detect fraud based on voice data, but do not take into account the user's psychological state, making it difficult to make accurate judgments. Since a rapid response to new fraud methods is also necessary, the present invention provides a system that solves these problems.

[0969] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving voice input, means for transmitting the received voice data to the server, means for converting the voice data to text data at the server, means for detecting specific phrases in the converted text data, means for evaluating the risk of the detected specific phrases, means for generating an alert based on the evaluation result, means for notifying pre-registered contacts of the generated alert, means for analyzing the user's emotional state, and means for reflecting the analyzed emotional state in a fraud risk evaluation. This enables highly accurate fraud detection and rapid response that takes the user's psychological state into consideration.

[0970] A "means for receiving voice input" is a device or system that converts a user's speech into an electrical signal and receives it in digital form.

[0971] The "means for transmitting received voice data to a server" refers to a method or protocol for transmitting voice data acquired by a terminal to a server via the Internet or other networks.

[0972] "Means for converting voice data into text data at the server" refers to the technology or algorithm that analyzes received voice data and converts the content into text format.

[0973] The "means for detecting specific phrases within the converted text data" refers to a filtering technique for identifying predefined specific words or phrases from the text data.

[0974] "Means for assessing the risk level of specific detected phrases" refers to algorithms or methods that assess the risk of the content of the call based on identified phrases and score them according to certain criteria.

[0975] The "means for generating alerts based on the evaluation results" is a system for creating warning messages and notifications based on the results of the risk evaluation.

[0976] "Means for notifying pre-registered contacts of generated alerts" refers to a mechanism for distributing generated warning messages to pre-defined contacts of next of kin and relevant authorities.

[0977] "Means for analyzing the user's emotional state" refers to technology that analyzes input data such as voice data and identifies the emotions (joy, anger, sadness, anxiety, etc.) felt by the user.

[0978] The "means for reflecting the analyzed emotional state in the fraud risk assessment" is a method for incorporating the user's emotional state into the risk assessment process and adjusting the risk score based on the results.

[0979] This invention combines an emotion engine with a system for preventing special frauds. Specifically, the system monitors users' calls in real time, assesses the possibility of fraud, and alerts relatives and relevant authorities if necessary.

[0980] The system is configured as follows:

[0981] Hardware and Software

[0982] The system is built using the user's smartphone, a server, and related software.

[0983] Hardware: Smartphone microphone, cellular network

[0984] Software: WebSocket (real-time transmission of voice data), requests (transmission of HTTP requests), speech-to-text API (conversion of voice data into text), emotion_recognition (analysis of emotional states), notification-service API (sending alerts)

[0985] Data processing and calculation

[0986] 1. Receiving and sending voice input

[0987] From the moment the user answers the call, the smartphone device collects the call audio in real time. Voice input is captured using a microphone, and then the audio data is sent to the server using WebSocket.

[0988] 2. Speech Recognition and Text Conversion

[0989] The server receives the voice data sent from the device. The received voice data is first converted into an appropriate format, and then converted into text data using a speech recognition engine. High-precision conversion from voice to text is performed, and the generated text data is then analyzed.

[0990] 3. Fraudulent word detection and evaluation

[0991] The server analyzes the text data generated by the speech recognition to detect specific fraud words and phrases using a predefined list of fraud-related keywords, which are then fed into a scoring algorithm to assess risk.

[0992] 4. Emotion Recognition by Emotion Engine

[0993] The server then inputs the received voice data into an emotion engine to analyze the user's emotional state in real time. The emotion engine analyzes changes in voice tone, speed, and volume to identify emotions such as joy, anger, sadness, and anxiety.

[0994] 5. Evaluations that reflect emotional states

[0995] The server will then evaluate the user's emotional state as determined by the emotion engine along with the detected fraud phrases. If the user is in a certain emotional state (e.g., anxiety or confusion), the likelihood of fraud increases, and points are added to the score.

[0996] 6. Alerting and Notification

[0997] Based on the results of the risk assessment, the server generates an alert, which includes a summary of the call, detected scam words, reputation score, text data, and the perceived emotional state. The alert is then sent to the user's contacts, including their next of kin and relevant authorities. Notifications can be sent via email, SMS, in-app notifications, and other methods.

[0998] 7. Always keep up with the latest criminal techniques

[0999] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[1000] Examples of prompts

[1001] For example, if a user includes a phrase such as "Please hurry up and transfer the money" during a call, while the emotion engine detects an emotional state such as anxiety or confusion, the smartphone device sends the voice data to the server in real time. The server converts the voice into text and detects the phrase "Please hurry up and transfer the money" and the user's sense of anxiety. The risk scoring algorithm evaluates these as high risk and immediately generates an alert to notify relatives and the police. This process makes it possible to prevent fraud before it occurs.

[1002] Example prompt sentence:

[1003] Create an application that immediately alerts relatives and the police if a user receives a phone call asking them to transfer money and feels uneasy. This application uses WebSocket to transmit voice data in real time and combines a speech recognition engine and an emotion recognition engine to evaluate fraudulent words and emotional state.

[1004] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1005] Step 1:

[1006] The terminal uses a microphone to collect call audio in real time from the moment the user answers the phone. The input is the user's speech, and the output is digitized audio data, which is sent to a server for further processing.

[1007] Step 2:

[1008] The device sends the collected audio data to the server using WebSocket. The input is the digitized audio data collected by the device, and the output is the audio data sent to the server in real time.

[1009] Step 3:

[1010] The server converts the received voice data into an appropriate format and inputs it to the voice recognition engine. The input is the voice data sent from the terminal, and the output is data in the appropriate format that is input to the voice recognition engine.

[1011] Step 4:

[1012] The server converts the voice data into text data using a speech recognition engine, where the input is the voice data converted into an appropriate format and the output is text data.

[1013] Step 5:

[1014] The server analyzes the text data generated by the speech recognition to detect specific fraudulent words and phrases. The input is the text data converted from speech, and the output is information containing the detected fraudulent words and phrases.

[1015] Step 6:

[1016] The server uses the detected fraud words and phrases to perform a risk assessment. The input is the detected fraud words and phrases, and the output is a fraud risk assessment score, which is used to generate subsequent alerts.

[1017] Step 7:

[1018] In parallel, the server inputs the voice data into the emotion engine to analyze the user's emotional state in real time. The input is the voice data received by the server, and the output is information identifying the user's emotional state (joy, anger, sadness, anxiety, etc.).

[1019] Step 8:

[1020] The server reflects the detected fraudulent words and the user's emotional state recognized by the emotion engine in the evaluation. The input is information on fraudulent words and emotional state, and the output is a detailed risk evaluation score that combines these.

[1021] Step 9:

[1022] Based on the risk assessment results, the server generates an alert, which includes a summary of the call, detected fraudulent words, the assessment score, text data, and the recognized emotional state. The input is the detailed risk assessment score and relevant data, and the output is the alert information.

[1023] Step 10:

[1024] The generated alerts are sent to pre-registered contacts of relatives and relevant authorities via email, SMS, in-app notifications, etc. The input is the generated alert information, and the output is the response of relatives and relevant authorities who receive the notification.

[1025] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1026] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1027] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1028] [Fourth embodiment]

[1029] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1030] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1031] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1032] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1033] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1034] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1035] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1036] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1037] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1038] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1039] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1040] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1041] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1042] This invention is a system for preventing special frauds. The system consists of terminal, server, and user components that work together to monitor and evaluate potentially fraudulent calls in real time, and alert relatives and relevant authorities as necessary.

[1043] Receiving and sending audio input

[1044] The device collects the call audio in real time from the moment the user answers the call. Audio input is captured using a microphone or similar audio input device, and the audio data is then transmitted to a server, typically using WebSocket or another real-time communication protocol.

[1045] Speech recognition and text conversion

[1046] The server receives the voice data sent from the device. The received voice data is converted into text data by a voice recognition engine. This voice recognition engine uses the latest machine learning technology, enabling highly accurate voice-to-text conversion.

[1047] Deceptive word detection and evaluation

[1048] The server analyzes the text data generated by the speech recognition to detect specific fraud words and phrases, typically using a predefined list of fraud-related keywords, which are then fed into a scoring algorithm to assess risk.

[1049] Alerting and Notifications

[1050] Based on the results of the risk assessment, the server generates an alert, which includes a summary of the call, detected scam words, and the text data of the call. The alert is then sent to pre-registered contacts, such as the user's relatives or relevant authorities. Notifications can be sent via email, SMS, in-app notifications, and other methods.

[1051] Always responding to the latest criminal techniques

[1052] The server continuously analyzes the received audio data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[1053] Specific examples

[1054] For example, if a user uses a phrase such as "Please hurry up and transfer the money" during a call, the device sends the voice data to the server in real time. The server converts the voice data into text and detects the phrase "Please hurry up and transfer the money." The risk scoring algorithm evaluates this phrase as high risk and immediately generates an alert to notify relatives and the police. This process can prevent fraud before it occurs.

[1055] This invention can effectively prevent damage caused by special frauds, and can significantly reduce the risk of being defrauded, especially for elderly people and single people. The specific programs and algorithms involved in implementing the invention will vary depending on the technology and environment to which it is applied, but they will generally be implemented based on the form shown here.

[1056] The processing flow will be explained below.

[1057] Step 1:

[1058] The terminal detects the call start event as soon as the user answers the phone and starts voice input. The voice is picked up in real time through a microphone and temporarily stored in a buffer as a digital signal.

[1059] Step 2:

[1060] The device transmits the captured audio data to the server in real time using WebSocket or other real-time communication protocols.

[1061] Step 3:

[1062] The server receives the voice data sent from the terminal, converts the received voice data into an appropriate format, and then inputs it into a voice recognition engine.

[1063] Step 4:

[1064] The server's speech recognition engine converts the speech data into text data. This engine performs highly accurate conversion from speech to text, and the generated text data is used for analysis.

[1065] Step 5:

[1066] The server analyzes the converted text data to detect specific fraudulent words and phrases using a pre-trained list of fraud-related keywords.

[1067] Step 6:

[1068] The server uses a scoring algorithm to assess the risk of detected scam words, with different weights assigned to different scam words to quantify their risk.

[1069] Step 7:

[1070] The server generates an alert based on the results of the scoring algorithm, which includes a summary of the call, detected scam words, reputation score, and text data.

[1071] Step 8:

[1072] The server then sends the generated alerts to pre-registered contacts of relatives and relevant authorities via email, SMS, in-app notifications, and more.

[1073] Step 9:

[1074] The server feeds the received voice and text data into a machine learning algorithm to constantly learn the latest fraud techniques, and the results of this learning are used for subsequent analyses.

[1075] Step 10:

[1076] The server stores all transaction logs in a database for future reference, including the start and end times of calls, analysis results, and notification content.

[1077] By going through the above steps, the present invention functions as an effective system for protecting users from special fraud.

[1078] Example 1

[1079] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1080] Currently, the methods of special fraud are evolving daily, increasing the risk of elderly people and single people falling victim to fraud. In particular, methods such as "bank transfer fraud" carried out over the phone are becoming more sophisticated, making it difficult for conventional prevention systems to deal with them. Therefore, a system is needed that analyzes the content of phone calls in real time, evaluates the risk of fraud, and issues alerts.

[1081] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1082] In this invention, the server includes a means for receiving voice input, a means for transmitting the received voice data to the server using a real-time communication protocol, and a means for converting the voice data into text data at the server, thereby enabling real-time analysis of the user's call content, detection of specific fraudulent words and phrases, and highly accurate risk assessment and prompt alert notification.

[1083] "Voice input" refers to a technology that uses a microphone or other voice input device to collect voices uttered by a user via a communication means such as a telephone.

[1084] A "real-time communication protocol" is a communication protocol for transmitting data such as telephone voice promptly with low latency. An example is WebSocket.

[1085] A "server" is a computer system that receives voice data sent from a terminal and performs processes such as voice recognition and data analysis.

[1086] "Text data" refers to character string information generated from voice data using voice recognition technology.

[1087] "Fraudulent words and phrases" are specific, predefined words or phrases related to fraudulent activity, such as "transfer" or "hurry."

[1088] A "scoring algorithm" is a calculation method used to numerically assess the risk of detected fraudulent words or phrases.

[1089] An "alert" is a warning message that is generated when dangerous call content is detected.

[1090] "Contacts" are people or organizations to whom alerts are sent. They are selected based on pre-registered information.

[1091] "Email" refers to digital letters sent and received over the Internet.

[1092] "SMS" is a means of sending and receiving short text messages over mobile phone networks.

[1093] "In-app notifications" refer to notification messages displayed through software such as smartphone apps.

[1094] The present invention is a system for preventing special frauds in advance, in which the components of the user, terminal, and server work together.

[1095] Receiving and sending audio input

[1096] The device collects the call audio in real time using a voice input device (e.g., a smartphone microphone) from the moment the user answers the call. The audio data is then sent to the server using a real-time communication protocol such as WebSocket.

[1097] Speech recognition and text conversion

[1098] The server receives the voice data sent from the device. The received voice data is converted into text data by a speech recognition engine (e.g., Google Cloud Speech-to-Text API). This speech recognition engine uses the latest machine learning technology, enabling highly accurate speech-to-text conversion.

[1099] Deceptive word detection and evaluation

[1100] The server analyzes the text data generated by the speech recognition. Natural Language Processing (NLP) algorithms are used for the analysis. The server uses a predefined list of fraud-related keywords to detect specific fraud words and phrases within the text data. The detected fraud words are then evaluated for risk using a scoring algorithm.

[1101] Alerting and Notifications

[1102] The server generates an alert based on the results of the risk assessment. The alert includes a summary of the call, detected scam words, and text data. The alert is then sent to pre-registered contacts, such as the user's relatives or relevant authorities, via email, SMS, or in-app notification.

[1103] Always responding to the latest criminal techniques

[1104] The server continuously analyzes the received audio data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[1105] Specific examples

[1106] For example, if a user uses a phrase such as "Please transfer the money quickly" during a call, the device will send the speech data to the server in real time. The server will convert the speech into text and detect the phrase "Please transfer the money quickly." The risk scoring algorithm will evaluate the phrase as high risk and immediately generate an alert to notify relatives and the police. This process can prevent fraud before it can be carried out.

[1107] Examples of prompt statements

[1108] "Detect specific phrases in the audio of a call and assess the likelihood of fraud. For example, if the phrase 'Please transfer the money quickly' is detected, issue a warning."

[1109] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1110] Step 1:

[1111] The device starts voice input.

[1112] The device uses a microphone to collect the call voice in real time from the moment the user answers the call. Specifically, the voice input software on the device runs in the background and collects the user's voice data.

[1113] Input: User's voice

[1114] Output: Collected audio data

[1115] Step 2:

[1116] Sending audio data to the server

[1117] The device sends the collected audio data to a server using a real-time communication protocol such as WebSocket, which requires stable communication with low latency.

[1118] Input: Collected audio data

[1119] Output: Audio data sent to the server

[1120] Step 3:

[1121] The server receives the audio data.

[1122] The server receives the voice data sent from the terminal in real time. The server process stores the received data and prepares it for subsequent voice recognition processing.

[1123] Input: Audio data sent from the device

[1124] Output: Archived audio data

[1125] Step 4:

[1126] Converting audio data to text

[1127] The server converts the voice data into text using a speech recognition engine, for example, the Google Cloud Speech-to-Text API, which ensures high accuracy in the process.

[1128] Input: Archived audio data

[1129] Output: Text data

[1130] Step 5:

[1131] Text data analysis

[1132] The server analyzes the text data generated by the speech recognition, using natural language processing (NLP) algorithms to detect fraudulent words and phrases, using a predefined list of fraudulent words.

[1133] Input: Text data

[1134] Output: Detected fraudulent words and phrases

[1135] Step 6:

[1136] Risk assessment

[1137] The server evaluates the detected fraudulent words and phrases using a scoring algorithm, which generates a numerical score indicating the likelihood of fraud.

[1138] Input: Detected scam words and phrases

[1139] Output: Risk score

[1140] Step 7:

[1141] Generate alerts

[1142] The server generates an alert based on the results of the risk assessment, which includes a summary of the call, detected scam words, and text data.

[1143] Input: Risk score, text data

[1144] Output: Alert

[1145] Step 8:

[1146] Alert Notification

[1147] Generated alerts are sent to pre-registered contacts via email, SMS, or in-app notifications.

[1148] Input: Alert

[1149] Output: Notified relatives and relevant authorities

[1150] Step 9:

[1151] Learning new fraud methods

[1152] The server uses the received voice data and analysis results to automatically learn new fraudulent techniques through machine learning algorithms, allowing the system to adapt to the latest fraud techniques.

[1153] Input: Received voice data, analysis results

[1154] Output: Updated deceptive wordlist and scoring algorithm

[1155] (Application example 1)

[1156] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1157] In modern society, special frauds are on the rise, with victims becoming increasingly serious, especially among the elderly. Telephone frauds are particularly sophisticated, leaving many victims vulnerable to deception. Current technology does not provide sufficient means to detect and prevent these frauds in real time. Therefore, a system is needed that analyzes the content of phone calls in real time and immediately sends an alert if there is a possibility of fraud. The system itself must also be able to learn and evolve so that it can constantly respond to the latest fraud techniques.

[1158] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1159] In this invention, the server includes a means for receiving voice input, a means for transmitting the received voice data to the server, a means for converting the voice data to text data at the server, a means for detecting specific phrases in the converted text data, a means for evaluating the risk level of the detected specific phrases, a means for generating an alert based on the evaluation results, a means for notifying pre-registered contacts of the generated alert, a means for collecting voice data and detecting fraudulent keywords in real time, and a means for automatically generating an alert to notify contacts when fraudulent keywords are detected. This allows the content of a user's calls to be analyzed in real time, and an alert to be immediately generated if there is a high possibility of fraud, making it possible to prevent damage before it occurs. Furthermore, by analyzing received voice data and constantly learning the latest criminal techniques, the system continuously evolves and can respond to more advanced fraud techniques.

[1160] "Audio input" refers to the means by which an audio signal is input to a digital device.

[1161] "Device" means an electronic device used by an individual or organization for communication and data processing.

[1162] "Server" refers to a computer system that receives, stores, analyzes, and transmits data over a network.

[1163] "Audio data" refers to data that represents an audio signal in digital form.

[1164] "Text data" refers to data obtained by converting voice data into character information.

[1165] "Specific phrases" refer to specific words or expressions that indicate possible fraud.

[1166] "Real-time" refers to data collection, analysis, notification, etc. being carried out immediately without delay.

[1167] "Risk assessment" refers to the process of analyzing and assessing the risk level of specific detected phrases.

[1168] An "alert" is a message that sends a warning or notification when a specific condition is met.

[1169] "Pre-registered Contacts" refers to contact information that has been pre-registered in the system for the purpose of sending alerts.

[1170] "Scam words" refer to specific keywords or phrases that indicate potential fraud.

[1171] "Machine learning" refers to technologies and methods that allow computers to analyze large amounts of data and learn automatically through algorithms.

[1172] The system of the present invention is designed to prevent special frauds, and involves the cooperation of the terminal, server, and user components. First, when a user starts a call, the terminal collects the user's voice input in real time and transmits it to the server as voice data. This voice data is transmitted using WebSocket or other real-time communication protocols.

[1173] The server converts the received voice data into text data using a speech recognition engine. This speech recognition engine uses the latest machine learning technology, such as Google's Speech-to-Text API, to achieve highly accurate speech-to-text conversion. The converted text data is then sent to the scam word detection and evaluation process.

[1174] To detect fraudulent words, a predefined list of fraud-related keywords is used. The server analyzes the generated text data to check whether it contains any of the keywords in this list. If any of these keywords are detected, a scoring algorithm is applied to assess the risk level. This algorithm evaluates the risk level of the detected keywords and generates an alert based on the results.

[1175] The generated alert will be sent to pre-registered contacts via email, SMS, or in-app notification, including a summary of the call and a recording. The alert will include information on the call that is likely to be fraudulent, allowing the user's relatives or relevant authorities to take immediate action. Notifications can be sent via notification services such as Amazon SNS.

[1176] Furthermore, the system uses machine learning algorithms to continuously analyze the received voice data in order to constantly adapt to the latest criminal techniques. This allows the system to respond quickly to new fraud methods. When a user actually experiences a call containing fraudulent words, the system accumulates that experience as learning data to improve future detection accuracy.

[1177] For example, if a user makes a call containing the phrase "Please hurry up and transfer the money," the device will send the voice data to the server in real time. The server will convert the voice data into text and detect the phrase "Please hurry up and transfer the money." Based on this detection result, a risk assessment will be performed and if the call is deemed high risk, an alert will be generated immediately and notified to relatives and relevant authorities. This process can prevent fraud from being carried out.

[1178] An example of a prompt sentence is, "There is a risk of important information being leaked during the phone call. Please evaluate the phrase 'Please transfer the money quickly' as a possible fraud and generate an alert."

[1179] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1180] Step 1:

[1181] When a user receives a call, the device collects the voice input through a microphone. This voice input is recorded as digital voice data. The input data is processed in real time and then sent to the next step. The output is the captured voice data.

[1182] Step 2:

[1183] The device sends the collected audio data to the server over the network. This process uses WebSocket or other real-time communication protocols. The input is the audio data collected on the device, and the output is the audio data sent to the server.

[1184] Step 3:

[1185] The server converts the received voice data into text data using a voice recognition engine, such as Google's Speech-to-Text API. The input is the voice data sent to the server, and the output is the converted text data.

[1186] Step 4:

[1187] The server analyzes the generated text data and detects fraudulent words using a predefined list of fraud-related keywords. It checks whether the text data contains any of the keywords in this list. The input is the text data, and the output is a list of detected fraudulent words.

[1188] Step 5:

[1189] The server applies a scoring algorithm to assess the risk of detected fraudulent words. The algorithm assesses the risk level of the detected keywords and outputs the result. The input is a list of fraudulent words, and the output is a risk assessment score.

[1190] Step 6:

[1191] Based on the assessment results, the server generates an alert, which is accompanied by information such as a summary of the call and the recording. The input is the risk assessment score and text data, and the output is the alert message.

[1192] Step 7:

[1193] Generated alerts are sent to pre-registered contacts via email, SMS, or in-app notification. This is done by the server using a notification protocol. The input is the alert message and the output is the notified alert.

[1194] Step 8:

[1195] To stay up to date with the latest criminal techniques, the server continuously analyzes the received audio data and uses machine learning algorithms to learn new fraud techniques. The input is previously processed audio data, and the output is an updated fraud wordlist and rating model.

[1196] This can prevent users from being involved in fraudulent activities and improve the accuracy of the system.

[1197] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1198] This invention combines an emotion engine with a system for preventing special frauds. The system's components, including the terminal, server, and user, work together to monitor and evaluate potentially fraudulent calls in real time, alerting relatives and relevant authorities as necessary. It also simultaneously recognizes the user's emotional state and incorporates that information into the evaluation, enabling more accurate fraud prevention.

[1199] Receiving and sending audio input

[1200] The device collects the call audio in real time from the moment the user answers the call. Audio input is captured using a microphone or similar audio input device, and the audio data is then transmitted to a server, typically using WebSocket or another real-time communication protocol.

[1201] Speech recognition and text conversion

[1202] The server receives the voice data sent from the device. The received voice data is first converted into an appropriate format and then input into a voice recognition engine. The voice recognition engine converts the voice data into text data. This engine performs highly accurate conversion from voice to text and the generated text data is used for analysis.

[1203] Deceptive word detection and evaluation

[1204] The server analyzes the text data generated by the speech recognition to detect specific fraud words and phrases using a predefined list of fraud-related keywords, which are then fed into a scoring algorithm to assess risk.

[1205] Emotion recognition by emotion engine

[1206] The server then inputs the received voice data into an emotion engine to analyze the user's emotional state in real time. The emotion engine analyzes changes in voice tone, speed, and volume to identify emotions such as joy, anger, sadness, and anxiety.

[1207] Evaluations that reflect emotional states

[1208] The server then evaluates the detected fraudulent phrases along with the user's emotional state as recognized by the emotion engine. If the user is in a certain emotional state (e.g., anxious or confused), points are added to the score to increase the likelihood of fraud.

[1209] Alerting and Notifications

[1210] Based on the results of the risk assessment, the server generates an alert, which includes a summary of the call, detected scam words, reputation score, text data, and the perceived emotional state. The alert is then sent to the user's contacts, including their next of kin and relevant authorities. Notifications can be sent via email, SMS, in-app notifications, and other methods.

[1211] Always responding to the latest criminal techniques

[1212] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[1213] Specific examples

[1214] For example, if a user uses a phrase like "Please hurry up and transfer the money" during a call, while the emotion engine detects an emotional state such as anxiety or confusion, the device sends the voice data to the server in real time. The server converts the voice into text and detects the phrase "Please hurry up and transfer the money" and the user's sense of anxiety. The risk scoring algorithm evaluates these as high risk and immediately generates an alert to notify relatives and the police. This process can prevent fraud from being committed.

[1215] This invention can effectively prevent damage caused by special frauds, and can significantly reduce the risk of being defrauded, especially for elderly people and single people. The specific programs and algorithms involved in implementing the invention will vary depending on the technology and environment to which it is applied, but they will generally be implemented based on the form shown here.

[1216] The processing flow will be explained below.

[1217] Step 1:

[1218] When the user answers the phone, the terminal detects the call start event and starts voice input. The call voice is picked up through the microphone and stored in a buffer in real time as a digital signal.

[1219] Step 2:

[1220] The device transmits the captured audio data to the server in real time, typically using WebSocket or other real-time communication protocols.

[1221] Step 3:

[1222] The server receives the voice data sent from the device and converts it into an appropriate format, which is then sent to a voice recognition engine.

[1223] Step 4:

[1224] The server's speech recognition engine converts the speech data into text data, which is then analyzed.

[1225] Step 5:

[1226] The server analyzes the generated text data to detect specific fraudulent words and phrases using a predefined list of fraud-related keywords.

[1227] Step 6:

[1228] The server uses a scoring algorithm to assess the risk of detected fraudulent words, which assigns different weights to different fraudulent words and quantifies their risk.

[1229] Step 7:

[1230] In parallel, the server inputs the voice data into an emotion engine to analyze the user's emotional state in real time, analyzing parameters such as voice tone, speed, and volume to identify emotions.

[1231] Step 8:

[1232] The server evaluates the risk level by taking into account the detected fraudulent phrases and the user's emotional state as recognized by the emotion engine. For example, if the user is anxious or confused, an evaluation score is added.

[1233] Step 9:

[1234] Based on the scoring algorithm, the server generates an alert that includes a summary of the call, detected scam words, reputation score, text data, and the perceived emotional state.

[1235] Step 10:

[1236] The generated alerts are sent by the server to pre-registered contacts via email, SMS, in-app notifications, and other methods.

[1237] Step 11:

[1238] The server feeds the received voice and text data into a machine learning algorithm to constantly learn the latest fraud techniques, allowing it to respond to new fraud methods.

[1239] Step 12:

[1240] The server stores all processing logs in a database, including the start and end times of calls, analysis results, and notification content.

[1241] By going through these steps, the system can protect users from special fraud and provide a fast and effective response.

[1242] Example 2

[1243] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1244] Special fraud methods are becoming more sophisticated every year, and the elderly and people living alone are at particularly high risk of falling victim to fraud. Conventional fraud prevention systems have the problem of being unable to completely prevent fraud victims because they have difficulty accurately identifying the content of fraud and are unable to reflect the user's emotional state in their evaluation.

[1245] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for receiving voice input, means for transmitting the received voice data to the server, means for converting the voice data into text data at the server, means for detecting specific phrases in the converted text data, means for evaluating the risk level of the detected specific phrases, means for analyzing the received voice data and recognizing the user's emotional state in real time, means for reflecting the user's emotional state in the evaluation, means for generating an alert based on the evaluation result, and means for notifying the generated alert to a pre-registered contact. This enables the possibility of fraud to be detected with high accuracy and enables a prompt and accurate response that also takes into account the user's emotional state.

[1246] "Means for receiving audio input" refers to a microphone or similar audio capture device for capturing audio data, such as speech.

[1247] "Means for transmitting received voice data to a server" refers to a method for transferring collected voice data to a server via a communication network such as the Internet.

[1248] "Means for converting voice data into text data at the server" refers to the process of converting voice data into text format using a voice recognition engine.

[1249] "Means for detecting specific phrases within the converted text data" refers to algorithms that search and detect pre-defined fraud-related keywords and phrases within the text data.

[1250] "Means for assessing the risk of specific detected phrases" refers to a method for quantifying and assessing the likelihood of fraud or the risk based on the detected phrases.

[1251] "Means for analyzing received voice data and recognizing the user's emotional state in real time" refers to an emotion engine that analyzes voice data and identifies the user's emotional state in real time, such as anxiety, anger, joy, or sadness.

[1252] "Means for reflecting the user's emotional state in the assessment" refers to a method for integrating and reflecting the user's emotional state recognized by the emotion engine in the fraud risk assessment.

[1253] "Means for generating alerts based on assessment results" refers to the process for generating warning messages or notifications when a high likelihood of fraud is assessed.

[1254] "Means of notifying pre-registered contacts of generated alerts" refers to means of sending generated alerts to pre-designated contacts, such as next of kin or relevant authorities, via email, SMS, in-app notifications, etc.

[1255] The present invention combines an emotion engine with a system for preventing special frauds, and is realized by the cooperation of the terminal, server, and user components.

[1256] Receiving and sending audio input

[1257] The device detects the start of a call when the user makes or receives a call. It catches the incoming call event and recognizes the beginning of the call. When the user starts a call, the device collects audio data of the conversation in real time using the built-in microphone or an external audio capture device. This audio data is sent to the server using a real-time communication technology such as the WebSocket protocol. To minimize communication delays, the audio data is sent intermittently, often using buffering.

[1258] Speech recognition and text conversion

[1259] The server receives the voice data sent from the device. The received voice data is stored in a specific buffer, where noise removal and voice normalization are performed as a preparatory step for analysis. The processed voice data is then input into a speech recognition engine (e.g., Google Speech-to-Text or IBM Watson) to convert the voice data into highly accurate text data. This text data is then analyzed.

[1260] Deceptive word detection and evaluation

[1261] The server analyzes the text data generated by the speech recognition engine. The analysis tool searches for and detects predefined fraud-related keywords and phrases (e.g., "transfer" or "hurry") within this text data. The detected fraud words are then input into a risk scoring algorithm, which evaluates the likelihood of the fraud.

[1262] Emotion recognition by emotion engine

[1263] In parallel, the server inputs the received voice data into the emotion engine to analyze the user's emotional state in real time. The emotion engine analyzes changes in voice tone, speed, and volume to identify the user's emotional state (e.g., anger, sadness, joy, anxiety, etc.).

[1264] Evaluations that reflect emotional states

[1265] The server integrates the detected fraud phrases with the user's emotional state as recognized by the emotion engine to evaluate the overall likelihood of fraud. In particular, if the user is anxious or confused, these feelings are added to the evaluation score.

[1266] Alerting and Notifications

[1267] Based on the evaluation results, the server generates an alert, which includes a summary of the call, detected scam words, evaluation score, text data, and the recognized emotional state, and the alert is sent to the user's relatives and relevant authorities via email, SMS, in-app notification, etc.

[1268] Always responding to the latest criminal techniques

[1269] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraud tactics, constantly updating the system's fraud detection capabilities.

[1270] Specific examples

[1271] For example, if a user says "Please hurry up and transfer the money" during a call, and the emotion engine detects the user's anxiety, the device will send the voice data to the server in real time. The server will convert the voice data into text, detect the phrase "Please hurry up and transfer the money," and the emotion engine will confirm the user's anxiety. The risk scoring algorithm will evaluate these as high risk and immediately generate an alert to notify relatives and the police. This process can prevent fraud from being committed.

[1272] Prompt Sentence Examples

[1273] Example prompts for input to a generative AI model:

[1274] "Please explain in detail the risk assessment process when the phrase 'hurry and transfer money' and the user's state of anxiety are detected during a call."

[1275] "Please tell me specifically how to send real-time collected call audio to a fraud detection system."

[1276] This system can significantly reduce the risk of elderly people and single people falling victim to fraud.

[1277] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1278] Step 1:

[1279] A call begins when a user makes or receives a call. The terminal catches this call reception event and recognizes the beginning of the call. The input includes the call reception event by the user. The output is the recognition of the start of the call.

[1280] Step 2:

[1281] When a user initiates a call, the device collects real-time audio data from the conversation using a built-in microphone or an external audio capture device. The input includes the user's voice during the call. The output is the collected audio data.

[1282] Step 3:

[1283] The device transmits the collected audio data to the server using real-time communication technology such as the WebSocket protocol. To minimize communication delays, the audio data is transmitted intermittently using buffering. The input includes the collected audio data. The output is the audio data transmitted to the server.

[1284] Step 4:

[1285] The server receives the audio data sent from the device. The received audio data is stored in a specific buffer, and noise removal and audio normalization are performed as a pre-analysis step. The input includes the audio data sent from the device. The output is the pre-processed audio data.

[1286] Step 5:

[1287] The server inputs the processed speech data into a speech recognition engine (e.g., Google Speech-to-Text or IBM Watson) to convert the speech data into text format. The input includes the preprocessed speech data. The output is the generated text data.

[1288] Step 6:

[1289] The server receives the generated text data from the speech recognition engine and inputs it into an analysis tool to detect predefined fraud-related keywords and phrases. The input includes the generated text data. The output is the detected fraud words and phrases.

[1290] Step 7:

[1291] The server inputs the detected fraud words into a risk scoring algorithm to quantify the likelihood of fraud. The input includes the detected fraud words or phrases. The output is a risk score.

[1292] Step 8:

[1293] In parallel, the server inputs the collected voice data into the emotion engine to analyze the user's emotional state in real time. The input includes the collected voice data. The output is the identification result of the user's emotional state.

[1294] Step 9:

[1295] The server integrates the detected scam phrases with the user's emotional state as recognized by the emotion engine to assess the overall likelihood of fraud. If the user is anxious or confused, this assessment is taken into account. Inputs include scam words, risk scores, and the user's emotional state. The output is the integrated assessment result.

[1296] Step 10:

[1297] Based on the evaluation results, the server generates an alert, which includes a summary of the call, detected scam words, the evaluation score, text data, and the recognized emotional state. The input includes the integrated evaluation results. The output is the generated alert.

[1298] Step 11:

[1299] The server notifies the generated alert to the user's relatives and relevant authorities via email, SMS, in-app notification, etc. The input includes the generated alert. The output is the notification sent.

[1300] Step 12:

[1301] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraud techniques, thereby constantly updating the system's fraud detection capabilities. Inputs include stored voice and text data. Outputs are the results of learning new criminal techniques.

[1302] (Application example 2)

[1303] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1304] Special frauds are becoming more sophisticated every year, and many people, especially the elderly, are falling victim to them. Frauds committed during phone calls have limited reaction time, making it difficult to respond immediately. Conventional fraud prevention systems detect fraud based on voice data, but do not take into account the user's psychological state, making it difficult to make accurate judgments. Since a rapid response to new fraud methods is also necessary, the present invention provides a system that solves these problems.

[1305] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving voice input, means for transmitting the received voice data to the server, means for converting the voice data to text data at the server, means for detecting specific phrases in the converted text data, means for evaluating the risk of the detected specific phrases, means for generating an alert based on the evaluation result, means for notifying pre-registered contacts of the generated alert, means for analyzing the user's emotional state, and means for reflecting the analyzed emotional state in a fraud risk evaluation. This enables highly accurate fraud detection and rapid response that takes the user's psychological state into consideration.

[1306] A "means for receiving voice input" is a device or system that converts a user's speech into an electrical signal and receives it in digital form.

[1307] The "means for transmitting received voice data to a server" refers to a method or protocol for transmitting voice data acquired by a terminal to a server via the Internet or other networks.

[1308] "Means for converting voice data into text data at the server" refers to the technology or algorithm that analyzes received voice data and converts the content into text format.

[1309] The "means for detecting specific phrases within the converted text data" refers to a filtering technique for identifying predefined specific words or phrases from the text data.

[1310] "Means for assessing the risk level of specific detected phrases" refers to algorithms or methods that assess the risk of the content of the call based on identified phrases and score them according to certain criteria.

[1311] The "means for generating alerts based on the evaluation results" is a system for creating warning messages and notifications based on the results of the risk evaluation.

[1312] "Means for notifying pre-registered contacts of generated alerts" refers to a mechanism for distributing generated warning messages to pre-defined contacts of next of kin and relevant authorities.

[1313] "Means for analyzing the user's emotional state" refers to technology that analyzes input data such as voice data and identifies the emotions (joy, anger, sadness, anxiety, etc.) felt by the user.

[1314] The "means for reflecting the analyzed emotional state in the fraud risk assessment" is a method for incorporating the user's emotional state into the risk assessment process and adjusting the risk score based on the results.

[1315] This invention combines an emotion engine with a system for preventing special frauds. Specifically, the system monitors users' calls in real time, assesses the possibility of fraud, and alerts relatives and relevant authorities if necessary.

[1316] The system is configured as follows:

[1317] Hardware and Software

[1318] The system is built using the user's smartphone, a server, and related software.

[1319] Hardware: Smartphone microphone, cellular network

[1320] Software: WebSocket (real-time transmission of voice data), requests (transmission of HTTP requests), speech-to-text API (conversion of voice data into text), emotion_recognition (analysis of emotional states), notification-service API (sending alerts)

[1321] Data processing and calculation

[1322] 1. Receiving and sending voice input

[1323] From the moment the user answers the call, the smartphone device collects the call audio in real time. Voice input is captured using a microphone, and then the audio data is sent to the server using WebSocket.

[1324] 2. Speech Recognition and Text Conversion

[1325] The server receives the voice data sent from the device. The received voice data is first converted into an appropriate format, and then converted into text data using a speech recognition engine. High-precision conversion from voice to text is performed, and the generated text data is then analyzed.

[1326] 3. Fraudulent word detection and evaluation

[1327] The server analyzes the text data generated by the speech recognition to detect specific fraud words and phrases using a predefined list of fraud-related keywords, which are then fed into a scoring algorithm to assess risk.

[1328] 4. Emotion Recognition by Emotion Engine

[1329] The server then inputs the received voice data into an emotion engine to analyze the user's emotional state in real time. The emotion engine analyzes changes in voice tone, speed, and volume to identify emotions such as joy, anger, sadness, and anxiety.

[1330] 5. Evaluations that reflect emotional states

[1331] The server will then evaluate the user's emotional state as determined by the emotion engine along with the detected fraud phrases. If the user is in a certain emotional state (e.g., anxiety or confusion), the likelihood of fraud increases, and points are added to the score.

[1332] 6. Alerting and Notification

[1333] Based on the results of the risk assessment, the server generates an alert, which includes a summary of the call, detected scam words, reputation score, text data, and the perceived emotional state. The alert is then sent to the user's contacts, including their next of kin and relevant authorities. Notifications can be sent via email, SMS, in-app notifications, and other methods.

[1334] 7. Always keep up with the latest criminal techniques

[1335] The server continuously analyzes the received voice and text data and uses machine learning algorithms to learn the latest fraudulent techniques, allowing the system to constantly adapt to new fraudulent methods.

[1336] Examples of prompts

[1337] For example, if a user includes a phrase such as "Please hurry up and transfer the money" during a call, while the emotion engine detects an emotional state such as anxiety or confusion, the smartphone device sends the voice data to the server in real time. The server converts the voice into text and detects the phrase "Please hurry up and transfer the money" and the user's sense of anxiety. The risk scoring algorithm evaluates these as high risk and immediately generates an alert to notify relatives and the police. This process makes it possible to prevent fraud before it occurs.

[1338] Example prompt sentence:

[1339] Create an application that immediately alerts relatives and the police if a user receives a phone call asking them to transfer money and feels uneasy. This application uses WebSocket to transmit voice data in real time and combines a speech recognition engine and an emotion recognition engine to evaluate fraudulent words and emotional state.

[1340] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1341] Step 1:

[1342] The terminal uses a microphone to collect call audio in real time from the moment the user answers the phone. The input is the user's speech, and the output is digitized audio data, which is sent to a server for further processing.

[1343] Step 2:

[1344] The device sends the collected audio data to the server using WebSocket. The input is the digitized audio data collected by the device, and the output is the audio data sent to the server in real time.

[1345] Step 3:

[1346] The server converts the received voice data into an appropriate format and inputs it to the voice recognition engine. The input is the voice data sent from the terminal, and the output is data in the appropriate format that is input to the voice recognition engine.

[1347] Step 4:

[1348] The server converts the voice data into text data using a speech recognition engine, where the input is the voice data converted into an appropriate format and the output is text data.

[1349] Step 5:

[1350] The server analyzes the text data generated by the speech recognition to detect specific fraudulent words and phrases. The input is the text data converted from speech, and the output is information containing the detected fraudulent words and phrases.

[1351] Step 6:

[1352] The server uses the detected fraud words and phrases to perform a risk assessment. The input is the detected fraud words and phrases, and the output is a fraud risk assessment score, which is used to generate subsequent alerts.

[1353] Step 7:

[1354] In parallel, the server inputs the voice data into the emotion engine to analyze the user's emotional state in real time. The input is the voice data received by the server, and the output is information identifying the user's emotional state (joy, anger, sadness, anxiety, etc.).

[1355] Step 8:

[1356] The server reflects the detected fraudulent words and the user's emotional state recognized by the emotion engine in the evaluation. The input is information on fraudulent words and emotional state, and the output is a detailed risk evaluation score that combines these.

[1357] Step 9:

[1358] Based on the risk assessment results, the server generates an alert, which includes a summary of the call, detected fraudulent words, the assessment score, text data, and the recognized emotional state. The input is the detailed risk assessment score and relevant data, and the output is the alert information.

[1359] Step 10:

[1360] The generated alerts are sent to pre-registered contacts of relatives and relevant authorities via email, SMS, in-app notifications, etc. The input is the generated alert information, and the output is the response of relatives and relevant authorities who receive the notification.

[1361] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1362] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1363] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1364] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1365] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1366] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1367] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1368] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1369] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1370] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1371] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1372] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1373] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1374] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1375] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1376] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1377] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1378] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1379] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1380] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1381] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1382] The following is further disclosed regarding the above embodiment.

[1383] (Claim 1)

[1384] means for receiving audio input;

[1385] means for transmitting the received voice data to a server;

[1386] A means for converting voice data into text data in a server;

[1387] means for detecting specific phrases within the converted text data;

[1388] a means for assessing the riskiness of a particular detected phrase;

[1389] means for generating an alert based on the evaluation results;

[1390] a means of notifying pre-registered contacts of generated alerts;

[1391] A system including:

[1392] (Claim 2)

[1393] 2. The system according to claim 1, further comprising means for analyzing received voice data and constantly learning the latest criminal techniques.

[1394] (Claim 3)

[1395] 10. The system of claim 1, further comprising means for attaching information including a call summary and recording when generating the alert.

[1396] "Example 1"

[1397] (Claim 1)

[1398] means for receiving audio input;

[1399] means for transmitting the received voice data to a server using a real-time communication protocol;

[1400] A means for converting voice data into text data in a server;

[1401] means for detecting specific fraudulent words or phrases within the converted text data;

[1402] a means for assessing the riskiness of detected fraudulent words and phrases using a scoring algorithm; and

[1403] means for generating an alert based on the evaluation results;

[1404] A means of notifying pre-registered contacts of generated alerts via email, SMS, or in-app notifications, among other means;

[1405] A system including:

[1406] (Claim 2)

[1407] 10. The system of claim 1, further comprising means for automatically learning the latest fraud techniques using the received voice data and the analysis results.

[1408] (Claim 3)

[1409] 10. The system of claim 1, further comprising: means for attaching information including a summary of the call, detected fraud words, and text data when generating the alert.

[1410] "Application Example 1"

[1411] (Claim 1)

[1412] means for receiving audio input;

[1413] means for transmitting the received voice data to a server;

[1414] A means for converting voice data into text data in a server;

[1415] means for detecting specific phrases within the converted text data;

[1416] a means for assessing the riskiness of a particular detected phrase;

[1417] means for generating an alert based on the evaluation results;

[1418] a means of notifying pre-registered contacts of generated alerts;

[1419] A means for collecting voice data and detecting fraudulent words in real time;

[1420] A means to automatically generate alerts to notify contacts when scam words are detected;

[1421] A system including:

[1422] (Claim 2)

[1423] 2. The system according to claim 1, further comprising means for analyzing received voice data and constantly learning the latest criminal techniques.

[1424] (Claim 3)

[1425] 10. The system of claim 1, further comprising means for attaching information including a call summary and recording when generating the alert.

[1426] "Example 2: Combining Emotion Engines"

[1427] (Claim 1)

[1428] means for receiving audio input;

[1429] means for transmitting the received voice data to a server;

[1430] A means for converting voice data into text data in a server;

[1431] means for detecting specific phrases within the converted text data;

[1432] a means for assessing the riskiness of a particular detected phrase;

[1433] means for analyzing the received voice data and recognizing the user's emotional state in real time;

[1434] a means for reflecting the user's emotional state in the evaluation;

[1435] means for generating an alert based on the evaluation results;

[1436] a means of notifying pre-registered contacts of generated alerts;

[1437] A system including:

[1438] (Claim 2)

[1439] 2. The system according to claim 1, further comprising means for analyzing received voice data and constantly learning the latest criminal techniques.

[1440] (Claim 3)

[1441] 10. The system of claim 1, further comprising means for attaching information including a call summary, a recording, and a recognized emotional state when generating the alert.

[1442] "Application example 2 when combining emotion engines"

[1443] (Claim 1)

[1444] means for receiving audio input;

[1445] means for transmitting the received voice data to a server;

[1446] A means for converting voice data into text data in a server;

[1447] means for detecting specific phrases within the converted text data;

[1448] a means for assessing the riskiness of a particular detected phrase;

[1449] means for generating an alert based on the evaluation results;

[1450] a means of notifying pre-registered contacts of generated alerts;

[1451] means for analyzing the emotional state of a user;

[1452] A means for reflecting the analyzed emotional state in a fraud risk assessment;

[1453] A system including:

[1454] (Claim 2)

[1455] 2. The system according to claim 1, further comprising means for analyzing received voice data and constantly learning the latest criminal techniques.

[1456] (Claim 3)

[1457] 10. The system of claim 1, further comprising means for attaching information including a call summary and recording when generating the alert. [Explanation of symbols]

[1458] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. means for receiving audio input; means for transmitting the received voice data to a server; A means for converting voice data into text data in a server; means for detecting specific phrases within the converted text data; a means for assessing the riskiness of a particular detected phrase; means for generating an alert based on the evaluation results; a means of notifying pre-registered contacts of generated alerts; A system including:

2. 2. The system according to claim 1, further comprising means for analyzing the received voice data and constantly learning the latest criminal techniques.

3. 10. The system of claim 1, further comprising means for attaching information including a call summary and recording when generating the alert.

Images