Apparatus and method for diagnosing dementia based on non-face-to-face treatment and inbound request

Abstract

The present invention relates to an apparatus and a method for performing a dementia test on the basis of non-face-to-face oral questions and answers. Dementia test request data and personal information are collected from user terminals, and administrative district and medical institution jurisdiction information is generated on the basis of originating location information. Accordingly, a person who wants a dementia test can conveniently perform a non-face-to-face dementia test at a desired time and place without having to visit a dementia test center or medical institution in person, thereby improving the convenience of the dementia test.

Description

Dementia diagnosis device and method based on non-face-to-face medical consultation and inbound requests

[0001] The present disclosure relates to a dementia testing device and method, and more specifically, to a non-face-to-face medical consultation and inbound request-based dementia testing device and method that initiates a telephone call capable of conducting a dementia test at the request of a person requiring a dementia diagnosis and performs a voice question-and-answer-based dementia test.

[0002] Alzheimer's disease (AD) is a brain disorder associated with aging, resulting from brain abnormalities that cause the progressive decline of memory. Furthermore, Alzheimer's disease can progress to dementia, which involves a persistent and overall decline in cognitive function severe enough to cause difficulties in daily life. Here, cognitive function refers to various intellectual abilities such as memory, language skills, spatial awareness, judgment, and abstract thinking, and each cognitive function is closely associated with specific areas of the brain.

[0003] Mild cognitive impairment (MCI) refers to a condition in which memory and cognitive function are impaired compared to the same age group, although it has not progressed to dementia. As MCI can develop into Alzheimer's disease, it is a crucial stage for early detection and preventive measures.

[0004] Such dementia (or mild cognitive impairment) screenings generally proceed in the order of screening, diagnostic, and differential diagnosis. Screening is conducted at local dementia care centers and public health centers; however, there are issues regarding the validity and efficiency of the tests because they require in-person visits by subjects and are performed manually. Specifically, the validity of the tests is diminished because individuals who can visit the centers in person are highly unlikely to be dementia patients, and significant losses in terms of time and cost occur as it takes tens of minutes to test a single person when performed manually.

[0005] Dementia screening via telephone is a phone-based dementia testing technology utilizing artificial intelligence, in which AI conducts the conversation with the test subject instead of a human.

[0006] The problem that the present disclosure aims to solve is to provide a device and method that allow a person who wishes to be diagnosed with dementia to conveniently perform a dementia test remotely at a time and place of their choosing, without the need to visit a dementia testing center or medical institution in person.

[0007] Furthermore, the problem that the present disclosure aims to solve is to provide an automated system that enables follow-up care for potential dementia patients by enhancing connectivity between dementia screening centers or medical institutions.

[0008] According to one aspect of the present disclosure, a dementia testing device providing a non-face-to-face voice question-and-answer-based dementia testing process comprises: an input module that collects first data including a user's request for a dementia test through at least one method among telephone calling, code scanning, web address input, text message transmission, app push touch, NFC tag contact, and kiosk input; a communication module that communicates with an external device to provide the dementia testing process, transmits a question voice to the external device, receives an answer voice from the external device, and transmits a result by the dementia testing process; a memory in which the voice question-and-answer-based dementia testing process is stored; and a processor that performs operations according to the process, wherein the processor collects second data including at least one of the user's name, phone number, gender, and highest level of education, and consent information for the use of personal information regarding the use of the second data together with the first data through the input module, analyzes information regarding the external device from which the first data was generated and transmitted, and configures a dementia testing process tailored to the user's environment based on the information regarding the external device, the first data, and the second data.

[0009] Additionally, according to one aspect of the present disclosure, a non-face-to-face inbound request-based dementia testing method comprises the steps of: collecting first data containing a user's request for dementia testing by at least one of telephone calling, code scanning, web address input, text messaging, app push touch, NFC tag contact, and kiosk input; collecting second data containing at least one of the user's name, phone number, gender, and highest level of education; and collecting consent information for the use of personal information regarding the use of said second data; identifying location information of an external device at the time of transmission of the first data; generating first jurisdiction information regarding the administrative district where said external device is located and second jurisdiction information regarding a medical institution responsible for said location based on the identified location information; subdividing said second jurisdiction information into metropolitan medical institutions and basic medical institutions; and adding guidance for each medical institution and an additional testing process to the voice question-and-answer-based dementia testing process basically provided to said external device according to said second jurisdiction information. and may include the step of transmitting a question voice to the external device, receiving an answer voice from the external device, calculating a score based on the dementia test process, and transmitting the calculated score and the result to the external device.

[0010] Other specific details of the present disclosure are included in the detailed description and drawings.

[0011] According to the present disclosure, a person who wishes to undergo a dementia test can conveniently perform the test remotely at a time and place of their choosing without having to visit a dementia test center or medical institution in person, thereby improving the convenience of the dementia test.

[0012] Since the process is conducted via an inbound method where the subject requests the test first, there is no need to wait for a call regarding a potential dementia test, thereby enhancing the immediacy of the test.

[0013] In addition, according to the present disclosure, follow-up care for potential dementia patients can be provided, thereby saving costs such as the time required for dementia testing and labor costs.

[0014] FIG. 1 is a conceptual diagram of a dementia testing device according to one embodiment of the present disclosure.

[0015] FIG. 2 is an exemplary diagram regarding the process of a dementia test being performed by a dementia test device according to one embodiment of the present disclosure.

[0016] FIG. 3 is a conceptual diagram regarding an inbound request of a dementia testing device according to one embodiment of the present disclosure.

[0017] FIG. 4 is a flowchart schematically illustrating a part of an inbound request-based dementia test method according to one embodiment of the present disclosure.

[0018] FIG. 5 is an example diagram regarding the addition of additional content using location information of an external device in a dementia test process according to one embodiment of the present disclosure.

[0019] FIG. 6a is a conceptual diagram showing the time-series process of a dementia test process according to one embodiment of the present disclosure.

[0020] FIG. 6b is a conceptual diagram showing the time-series process of a dementia test process with additional content added using location information of an external device according to one embodiment of the present disclosure.

[0021] FIG. 7 is an exemplary diagram relating to an STT multiplexing process in a dementia testing device according to one embodiment of the present disclosure.

[0022] FIG. 8 is a flowchart schematically illustrating a part of a dementia test method according to one embodiment of the present disclosure.

[0023] The advantages and features of the present disclosure and the methods for achieving them will become clear by referring to the embodiments described below in detail together with the accompanying drawings. However, the present disclosure is not limited to the embodiments disclosed below but may be implemented in various different forms. These embodiments are provided merely to ensure that the disclosure of the present disclosure is complete and to fully inform those skilled in the art of the scope of the present disclosure, and the present disclosure is defined only by the categories of the claims.

[0024] The terms used in this specification are for describing embodiments and are not intended to limit the disclosure. In this specification, the singular form includes the plural form unless specifically stated otherwise in the text. The terms “comprises” and / or “comprising” as used in this specification do not exclude the presence or addition of one or more other components in addition to the components mentioned. Throughout the specification, the same reference numerals refer to the same components, and “and / or” includes each of the mentioned components and all combinations of one or more. Although terms such as “first,” “second,” etc., are used to describe various components, these components are not limited by these terms. These terms are used merely to distinguish one component from another. Accordingly, the first component mentioned below may be the second component within the technical scope of this disclosure.

[0025] In this specification, "Artificial intelligence call" refers to a telephone equipped with artificial intelligence technology, meaning a technology in which AI performs the call instead of a human. The AI ​​call system converts the user's voice into text data using Speech-to-Text (STT) technology and converts the text into speech using Text-to-Speech (TTS) technology. The converted text data is analyzed through Natural Language Processing (NLP) technology to identify the user's intent and generate appropriate responses. In the field of dementia screening, AI calls can be used to automate screening tests and determine whether follow-up tests are necessary. This allows subjects to receive tests via telephone without having to visit a testing institution in person, and reduces testing time and costs by automating the process of human examiners asking questions and recording answers. AI calls can also assist in identifying subjects requiring follow-up tests by analyzing test results.

[0026] In this specification, "dementia test" refers to a non-face-to-face dementia test utilizing an artificial intelligence call system. Since the AI ​​call system performs the dementia test via telephone, the test subject does not need to visit the testing institution in person, and it shortens testing time by automating the questioning, answer recording, and analysis processes performed by a human examiner. It can operate 24 hours a day and process multiple tests simultaneously, allowing for the efficient utilization of medical resources. It expands testing opportunities for the elderly who have difficulty moving or lack time, and enables testing to be provided without geographical restrictions. Artificial intelligence can perform objective evaluations by quantitatively analyzing various linguistic characteristics such as word count, conversion score, category score, and perseveration score. Dementia tests based on the AI ​​call system can analyze test results to identify subjects requiring follow-up testing and connect them to specialized medical institutions or dementia testing centers.

[0027] In this specification, "verbal fluency value" refers to a numerical value of related abilities (semantic memory, executive function, working memory, etc.) to determine the presence and / or degree of progression of dementia. It plays an important role in determining the likelihood of dementia by quantitatively evaluating the language ability of a test subject. Generally, as dementia patients experience a decline in language ability, 1) the number of words they can speak within a given time decreases, 2) they tend to repeat the same words, and 3) they frequently deviate from the topic of conversation or provide answers unrelated to the question. Therefore, to obtain the "verbal fluency value," a question is asked to speak words related to a specific topic, the audio response is converted into text data, and words corresponding to the given topic are extracted.

[0028] Scores are assigned based on the total number of words, the number of words in the first and second halves, the number of characters per word, the number of category changes, the number of words per category, and the number of duplicate words. The calculated language fluency value is compared with a pre-set threshold to determine whether a follow-up test is necessary. The threshold is established for each test group based on auxiliary information such as gender, age, education level, and number of cohabitants.

[0029] "Verbal fluency values" play an important role in AI call system-based dementia testing, and 1) shorten test time: increase efficiency by reducing test time to within 3 minutes 2) minimize target restrictions: since it is not a paper-and-pencil test, there are almost no restrictions on the subjects to be tested 3) non-face-to-face testing: the test can be conducted without the need for the examiner and the test subject to meet in person 4) since various instructions can be provided in the form of an app or web on a smartphone screen simultaneously with the call, accurate and convenient testing can be performed anytime and anywhere.

[0030] "Language fluency" can be measured by the following formula.

[0031] [Mathematical Formula 1]

[0032] LFV=a1*A+a2*B+a3*C+a4*D+a5*E+a6*F+a7*G

[0033] LFV: Language fluency value

[0034] A: Total number of words

[0035] B: Number of words in the first half

[0036] C: Number of words in the latter part

[0037] D: Average number of characters per word

[0038] E: Number of category changes

[0039] F: Average number of words per category

[0040] G: Number of duplicate words

[0041] a1~a7: Weight variables (satisfying a1+a2+a3+a4+a5+a6+a7=1, where a1~a7 are each constants between 0 and 1 inclusive)

[0042] In one embodiment, the language fluency value can be calculated by reflecting not only the content of the test subject's answer but also auxiliary information. Here, "auxiliary information" is basic information such as personal details of the test subject, and may include, for example, gender (male, female), age (teens, 20s, 30s, 40s, 50s, 60s, 70s, 80s, 90s, etc.), education level (elementary school graduate, middle school graduate, high school graduate, master's degree, bachelor's degree, doctorate, etc.), number of cohabitants (1 person, 2 people, 3 people, 4 people, etc.). The test content analysis unit (145) can calculate the language fluency value more accurately by differentially applying the types, number, weights, etc. of variables used to calculate the language fluency value based on auxiliary information such as age, gender, education level, and number of cohabitants.

[0043] Unless otherwise defined, all terms used herein (including technical and scientific terms) may be used in a meaning commonly understood by those skilled in the art to which this disclosure pertains. Additionally, terms defined in commonly used dictionaries are not to be interpreted ideally or excessively unless explicitly and specifically defined otherwise.

[0044] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the attached drawings.

[0045] FIG. 1 is a conceptual diagram of a dementia testing device according to one embodiment of the present disclosure.

[0046] As illustrated in FIG. 1, a dementia test device (100) according to one embodiment of the present disclosure can perform a dementia test by exchanging information with an external device (200) including a mobile device. The dementia test device (100) may include an input module (110), a dementia test module (120), a processor (130), a voice conversion module (140), a memory (150), a communication module (160), and an STT module (170).

[0047] The dementia test device (100) is a server that performs a dementia test using an artificial intelligence call. Here, "artificial intelligence call" is a telephone to which artificial intelligence technology is applied, and is a technology in which artificial intelligence performs the call instead of a person. As a specific example, the artificial intelligence call can provide a pre-stored voice consisting of an artificial intelligence voice and / or a voice recorded by a person to a terminal connected to the artificial intelligence call. For example, the pre-stored voice may include guidance messages, question voices, closing messages, etc.

[0048] The user can participate in the dementia test by interacting with the AI, such as by receiving voice provided by the AI ​​call and responding accordingly. Meanwhile, in one embodiment, if it is determined that the dementia test process is not proceeding smoothly with the AI ​​call-based user, it is possible to stop providing the pre-stored AI voice and for a person to directly intervene and communicate with the dementia test participant, who is the user of the external device (200). When the user thinks that the dementia test is not proceeding smoothly, they can manually contact the dementia test service provider by pressing the 'Call Counselor' button displayed on the screen of their external device (200).

[0049] The dementia testing device (100) can communicate with an external device (200). When the dementia testing device (100) establishes a phone call with the external device (20) of the user who is the subject of the test, it initiates an artificial intelligence call and performs a voice question-and-answer-based dementia test. During the test, the dementia testing device (100) analyzes the results of the dementia test using the voice answers obtained from the external device (200) as the test content to determine whether the user of the external device (20) is a subject for a subsequent dementia test. This process is performed automatically by the dementia testing device (100) without direct intervention by a 'person' such as an examiner or administrator, thereby significantly saving time and costs associated with the dementia test. Additionally, by filtering in advance whether a visit test is necessary via a phone call before the subject visits the testing institution to participate in the test, the effectiveness of the subsequent dementia test can be enhanced and the budget reduced.

[0050] The dementia testing device (100) evaluates the language fluency of a test subject through voice question-and-answer based on artificial intelligence call and determines whether a follow-up test is necessary. Specifically, the dementia testing device (100) conducts a voice question-and-answer-based dementia test and analyzes the test subject's response voice obtained therefrom to calculate a language fluency value that quantifies the language fluency. The dementia testing device (10) determines whether a follow-up test is necessary by comparing and analyzing the calculated language fluency value with a preset value. For example, if the language fluency value is less than the preset value, the user may be determined to be a subject for a follow-up test, and if the language fluency value is greater than or equal to the preset value, the user may be determined not to be a subject for a follow-up test. The dementia test performed by the dementia testing device (100) may correspond to any one of three stages of testing consisting of a screening test, a diagnostic test, and a differential test.

[0051] The input module (110) can collect first data containing a user's dementia test request in order to initiate a dementia test process through the dementia test device (100). Since the first data containing the dementia test request is generated by a user possessing an external device (200), if it is not properly controlled, the server may crash or the network may be paralyzed due to the occurrence of an excessive dementia test process. This leads to a waste of budget and can have a significant impact on the dementia management business. Therefore, the inbound dementia test device must handle the 'dementia test request' signal precisely.

[0052] The dementia test process of the inbound dementia test device (100) is initiated by first data including a request for a dementia test. The first data may be collected through one or more methods such as telephone outgoing, scanning a two-dimensional barcode, entering a web address, sending an SMS text message, touching an app push, contacting an NFC tag, or entering a kiosk. The first data is mostly generated by an external device (200), such as a smartphone owned by a user who wishes to participate in the dementia test, but may also be generated by a desktop PC, kiosk, or computer connected to an input device, rather than a smartphone, and transmitted to the dementia test device (100). The external device (200) may be a mobile device owned by the test subject, or a tablet PC, desktop PC, kiosk, or computer connected to an input device. However, the device on which the dementia test process is to proceed must be a communication device equipped with a microphone.

[0053] The first data containing the user's dementia test request can be generated (300) by one or more of the following methods: telephone call, two-dimensional barcode scanning, web address input, SMS text transmission, app push touch, NFC tag contact, and kiosk input. From the perspective of the dementia test device (100), the 'dementia test request' is received, so it is called an 'inbound request'. Referring to FIG. 3, the configuration of the inbound request (300) can be confirmed. The dementia test request generated by the external device (200) is transmitted to the dementia test device (100) through the input module (110) and serves as a trigger to start the dementia test process.

[0054] An example of the first data collected through a phone call is as follows. When a user makes a call to a designated phone number, the automatic response system (ARS) of the dementia test device (100) operates to collect a dementia test request. The automatic response system (ARS) is included in the input module (110). The dementia test device (100) recognizes the user's voice command and, if necessary, can induce the user to input basic information by voice, or allow the user to select by pressing a button if additional testing is required.

[0055] An example of the first data collected by scanning a two-dimensional barcode is as follows. When a user scans a QR code with a smartphone, a webpage or app opens, and a request for a dementia test is received by the dementia test device (100) through the input module (110). After receipt, a call can be automatically made to an external device (200) through the communication module (160) of the dementia test device (100), and when the call is connected, the dementia test process begins. In this process, the user's ID or initial information stored in the browser or smartphone is automatically entered via the QR code, allowing the dementia test to start quickly. It is not necessary to be a member, and the process can be conducted under a pseudonym. If the user consents on the browser called via the QR code, one or more of the second data, such as the user's name, phone number, gender, and highest level of education, may be collected. Additionally, consent information regarding the use of personal information concerning the use of the second data can be obtained from the user.

[0056] An example of the first data collected by entering a web address is as follows. A user can enter a pre-designated web address through a web browser pre-installed on a computer connected to the internet, such as a smartphone or PC, and move to a dementia test page to register a request. At this time, the web page is provided with a basic information input field so that the user can directly enter their name, contact information, phone number, desired test date, etc. Based on the entered information, the dementia test device (100) automatically makes a call to start the dementia test process.

[0057] In situations where the individual cannot personally perform the test, such as when parents above a certain age wish to undergo a dementia screening, test information can be pre-configured via a web address input. By notifying the subject in advance, family members can conveniently take the test. Furthermore, the pre-configured test information can be easily shared by sending it to the subject in advance via a messenger like KakaoTalk. Upon receiving the information, the subject can also adjust the schedule on the webpage to take the test at a time convenient for them. This web address input method has the advantage of easy access even without a QR code.

[0058] An example of the first data collected by SMS text transmission is as follows. When a user sends an SMS containing a specific code or keyword to a designated number, a request for a dementia test is automatically received. The dementia test device (100) analyzes the content of the text message sent by the user and the sender's number to identify the user who requires the test, and if it is determined that the user requires the test, attempts to initiate a phone call to start the dementia test process. If necessary, a detailed procedure or link for the test may be provided via a reply text message. If the desired date for the dementia test is specified in the content of the text message while sending the SMS message, the device analyzes the content of the message and can initiate a phone call for the dementia test on the date desired by the sender.

[0059] An example of the first data collected via an app push notification is as follows. If there is a dementia test app connected to the internet with the dementia test device (100), the user can initiate a test request via the app push notification. When the push notification is clicked, the user's identity information and basic status information are automatically transmitted to the app, and the test begins immediately. The method via the app allows users who have already installed the app to easily access it, and supports efficient testing by linking additional user information and records. Furthermore, since the user can check through the app when they received the dementia test, what the dementia test score was, and whether the dementia test score is improving, it can be efficient for long-term management of dementia patients in many ways. It is also convenient because it can easily provide information on lifestyle habits and dietary habits that are good for dementia, various games and content for dementia prevention, and the locations of dementia care centers and medical institutions.

[0060] An example of the first data collected by an NFC tag is as follows. When a user touches an external device (200), such as a smartphone, to an NFC tag at a specific location, a request for a dementia test is automatically collected. Recently, NFC tag technology has advanced, making it possible to simply install it inside thin paper media such as stickers or printing paper. Since most smartphones are equipped with a reader containing a sensor capable of recognizing NFC tags, a dementia test request signal can be generated by simply touching the smartphone to the NFC tag. In particular, since it is common to register the installation location when installing the NFC tag, it is easy to determine the user's location information even when using an external device (200) with location information such as GPS turned off. After determining the location information, a customized test can be provided by reflecting the starting location or situation of the dementia test.

[0061] An example of the first data collected by kiosk input is as follows. A user can directly request a dementia test through a kiosk installed in a hospital or public place. The kiosk is equipped with a touchscreen, allowing the user to press a request button, enter simple information, and apply for the test. The kiosk method can provide the function of requesting a test directly on-site and scheduling the dementia test at a time desired by the user. Therefore, a patient visiting the hospital for other medical treatment can conveniently schedule a dementia test, and an accompanying caregiver can schedule the test on their behalf.

[0062] The dementia test module (120) may be included in or connected to the processor (130) to receive and process data provided by the processor (130) and provide the processed value to the processor (130). The dementia test module (120) performs the role of conducting a dementia test on a test subject who possesses an external device (200) connected to an artificial intelligence call for the dementia test. The dementia test performed by the dementia test module (120) may be conducted based on voice question-and-answer. For example, the dementia test module (120) may conduct the dementia test by providing a question for the test to the external device (200) and obtaining the test subject's answer voice from the external device (200). Specifically, the dementia test module (120) may provide a question voice that provides a specific topic (or criteria, category, etc.) to the test subject and requests that the subject answer as many words corresponding to the topic as possible within a preset time. The dementia test module (120) may obtain the answer voice in which the user answers in response to the question voice.

[0063] In one embodiment, the dementia test performed by the dementia test module (120) may include a first test and a second test that are performed sequentially. Here, the “first test” may be a practice test performed prior to the second test to enhance the user’s understanding of the test. Additionally, the “second test” may be the main test that is substantially used to determine the presence or absence of dementia or the level of dementia symptoms of the test subject.

[0064] Detailed information regarding the dementia test performed by this dementia test module (120) will be described later with reference to FIGS. 2 to 4.

[0065] The processor (130) may be composed of one or more processors. The processor (130) can process data required for dementia testing by exchanging information with the input module (110), the dementia test module (120), the processor (130), the speech conversion module (140), the memory (150), the communication module (160), and the STT module (170). At this time, the one or more processors may be general-purpose processors such as CPUs, APs, and DSPs (Digital Signal Processors), graphics-dedicated processors such as GPUs and VPUs (Vision Processing Units), or artificial intelligence-dedicated processors such as NPUs. The one or more processors control the processing of input data according to predefined operation rules or artificial intelligence models stored in memory.

[0066] The predefined operation rules or artificial intelligence models are characterized by being created through learning. Here, being created through learning means that a predefined operation rules or artificial intelligence models configured to perform a desired characteristic (or objective) are created by a basic artificial intelligence model being trained using a number of training data by a learning algorithm. Such learning may be performed on the device itself where the artificial intelligence according to the present disclosure is executed, or it may be performed through a separate server and / or system. Examples of learning algorithms include supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning, but are not limited to the examples described above.

[0067] The processor (130) collects and processes second data containing one or more of the user's name, phone number, gender, and highest level of education, and personal information consent information regarding the use of said second data, along with the user's request for a dementia test, through an input module (110). In addition to the user's name, phone number, gender, and highest level of education, the second data may further include date of birth, address, phone number, family history, current disease status, medication information, work history, lifestyle information, history related to cognitive function, language usage habits, level of participation in social activities, and psychological state. A separate survey or game may be provided to collect information on said second data. To collect said second data without resistance, a reward may be given for proceeding with said survey or game.

[0068] Although the collection of secondary data is important to obtain more accurate dementia test results, since this information often constitutes personal information, it can be carried out by obtaining consent for personal information during inbound request processes such as making phone calls, scanning 2D barcodes, entering web addresses, sending SMS messages, touching app pushes, contacting NFC tags, and kiosk input.

[0069] The processor (130) can analyze information about the external device (200) to which the first data is generated and transmitted, and can configure a dementia test process tailored to the user's environment based on the external device information, the first data, and the second data. An example of the external device (200) to which the first data including a dementia test request is transmitted may be a mobile device such as a smartphone. After obtaining consent for personal information, the dementia test device (100) can collect location information of the external device (200), and can check whether there is any change in location at specific times or the frequency of outings of the user whose location information is activated.

[0070] If there is little change in location, it is determined that mobility is impaired, and the difficulty level of dementia tests related to cognitive function can be adjusted; if there is frequent movement, dementia tests related to spatial awareness can be included. Additionally, by analyzing patterns such as the language or words frequently used by the user, typing speed, and voice speed during calls, language usage habits can be identified based on language patterns, and the difficulty level of dementia tests can be increased or decreased based on frequently used words and context.

[0071] Furthermore, by analyzing whether a user frequently uses a smartphone or primarily uses it during specific time periods, the system can avoid those times or direct AI dementia test calls specifically during those times. Additionally, the type of dementia test can be selectively provided by analyzing the user's app usage patterns. By analyzing which apps—such as news, social media, messengers, or game apps—are primarily used, the system can conduct the dementia test process based on current affairs topics for users who mainly use news apps, or provide a test process on topics unrelated to current affairs. For users who primarily use social media or messenger apps, the difficulty of the test can be adjusted by conducting the process based on vocabulary frequently used or topics of interest expressed in those platforms, or by conducting the test on topics unrelated to those areas.

[0072] The processor (130) can identify the location information of the external device (200) at the time of transmission of the first data, and based on the identified location information, generate first jurisdiction information regarding the administrative district where the external device (200) is located and second jurisdiction information regarding the medical institution responsible for the location.

[0073] The processor (130) subdivides the second jurisdiction information into a wide-area medical institution and a basic medical institution, and according to the second jurisdiction information, can add guidance for each medical institution and an additional examination process to the voice question-and-answer-based dementia examination process provided to the external device (200) as a basic requirement.

[0074] The processor (130) subdivides into metropolitan administrative agencies and basic administrative agencies based on the first jurisdiction information, analyzes the number of dementia test requests, request method, request location, and request time of the user based on the first data, the second data, and the first jurisdiction information, and can process so as not to proceed with the dementia test process in the case of more than the number of dementia test requests within a pre-set period by the same user.

[0075] Detailed information regarding the identification of location information of an external device (200) and additional information processing based thereon in the dementia test process performed by such a processor (130) will be described later with reference to FIGS. 5 to 6a and 6b.

[0076] The processor (130) analyzes the response voice received through the communication module (160) in real time to separate the user's voice and background voice, and if the background voice includes the voice of a person other than the user or if it is determined that the background voice was generated in a place where there is noise of 50 decibels or more, it can transmit to the external device (200) whether to stop the dementia test process.

[0077] To separate user voices from background voices in real time, separation using microphone arrays and neural network-based voice separation models can be utilized. In the case of separation using microphone arrays, multiple microphone arrays are used to separate voice sources. Based on the location information of the sound sources, sounds originating from a specific direction can be emphasized, while sounds from other directions can be suppressed. For example, a user's voice originating close to the smartphone microphone can be emphasized, while background sounds or voices can be suppressed. Additionally, beamforming technology can be used to extract only the user's voice signal from a specific location, separating noise or the voices of others from other directions as background voices.

[0078] Neural network-based speech separation utilizes Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), or Transformer models to process speech data in real time and separate user voices from background voices. Using pre-trained speech separation models, techniques can be applied to isolate the user's voice even in situations where multiple speech signals are mixed. This allows for the real-time processing of response voice signals to separate background noise and the user's voice into independent signals. Signal-for-signal correlation analysis can also be applied to separate the user's response from background noise by analyzing the frequency bands of the speech signals to distinguish in real-time the characteristics of the speaker's voice (e.g., pitch, pronunciation patterns) and the characteristics of background noise (e.g., irregular frequency bands).

[0079] The method for determining whether background audio includes the voice of a person other than the user who is the subject of the dementia test is as follows. To distinguish the speaker in the voice signal, a speaker recognition algorithm is applied. After training the system with the user's voice in advance, it is possible to determine in real-time whether the voice in the response audio belongs to the same speaker. In the case of the first test, since it is a practice dementia test, the speaker recognition algorithm can be applied to recognize the voice of the user who is the subject of the dementia test while conducting the first test, and to distinguish between the voice of the user who is the subject of the dementia test and that of a person who is not during the second test, which is the main test.

[0080] In addition, while conducting the first test, the volume, timbre, and tone of the answer voice uttered by the user can be analyzed to distinguish whether the user is a subject for dementia testing. In this way, the characteristics of the voice signal for each speaker can be analyzed to determine the frequency and tone of the utterance. For example, by configuring the user's voice characteristics and the background voice characteristics as training data, the user's voice and the voice signal of another speaker can be separated. If it is determined that the background voice includes the voice of a person other than the user, the user can be notified of whether to stop the dementia test process in order to obtain an accurate dementia test result by transmitting to the external device (200).

[0081] A method for determining whether background voice is generated in a place where noise of 50 decibels or more exists can be a noise intensity analysis method through decibel measurement. The noise intensity in decibel (dB) units is measured from the answer voice signal received by the communication module (160). To do this, a noise level meter function is implemented to calculate the average dB value of background noise collected over a certain period of time, and if the calculated noise level exceeds 50 decibels, the noise generated in that place is determined to be background noise affecting the user's voice signal. It can also be done using a frequency band analysis method.

[0082] Generally, background noise appears in a specific frequency band, so the band is analyzed in real time to set a frequency range where noise of 50 decibels or more mainly occurs. If the signal strength in the specific frequency band is 50 decibels or more, it is determined that the background noise is strong and can be recognized as noise that may affect the user's response. This determination can also be made using noise removal filtering technology and level monitoring. To remove background noise from the voice signal, a noise cancelling filter is applied to extract only the noise components, and the decibels of the filtered background noise signal are monitored to determine if there is noise of 50 decibels or more. If it is determined that the background voice was generated in a place where noise of 50 decibels or more exists, a decision on whether to stop the dementia test process is transmitted to the external device (200), thereby inducing the dementia test to be conducted in an environment where the user can obtain more accurate dementia test results.

[0083] The processor (130) may include a dementia test module (120) that receives first data from the input module (110), connects a telephone call with an external device (200) for a dementia test, provides guidance voice through a communication module (160), conducts a first test which is a preliminary practice test to improve the subject's understanding of the test procedure, and conducts a second test which is a main test used to determine the presence or absence of dementia disease or the level of dementia symptoms of the subject.

[0084] The processor (130) can analyze the above-mentioned answer voice for the number of words, the number of words in the first half, the number of words in the second half, the number of characters per word, the number of category changes, the number of words per category, the speaking speed, pronunciation accuracy, intonation, stress change, voice tremor, topic consistency, maintenance of conversational flow, and the number of duplicate words, and apply a scoring criterion for at least one of these to score the result of the above-mentioned dementia test process.

[0085] The processor (130) controls an STT module (170) composed of one or more Speech-To-Text (STT) models, applies three or more STT models to convert the answer speech into text, compares the converted text words, selects the converted text that accounts for the majority of the converted text, and uses it to score the results of the dementia test process. The word processing process of the STT module (170) is illustrated in detail in FIG. 7.

[0086] The voice conversion module (140) can provide a customized experience to the user by utilizing various voice characteristics and can be converted into various voices to help interact with the user. 1) Voice conversion according to age is possible. The AI ​​voice can be converted according to age groups, such as young children, young adults, middle-aged people, and the elderly, to enable more natural and friendly communication with the user. 2) Voice conversion according to gender is possible. It can be converted into male and female voices to match user preferences. 3) Voice conversion according to emotion is possible.

[0087] Depending on the situation, the voice can be adjusted to sound calm, friendly, or encouraging. A calm tone is used during questioning phases where the user needs to concentrate, and an encouraging closing remark is provided when the diagnosis is complete to offer psychological stability. 4) Speed ​​and intonation can be adjusted. To enhance the effectiveness of information delivery, difficult questions can be pronounced slowly and clearly, while simple instructions can be delivered naturally and quickly. 5) It is possible to use specific regional dialects.

[0088] It is possible to enable more natural interaction by providing regional dialects or familiar accents tailored to the user's linguistic background. Intimacy can be formed by providing questions or guidance messages in the corresponding dialect to users from a specific region. In particular, by using a voice conversion module (140), not only questions but also the voice of answers provided by the dementia diagnosis subject can be converted into standard language, thereby obtaining more accurate dementia test results. A regional dialect database is constructed to analyze the answers, and dialects that have the same meaning as standard language are accepted as correct answers. Dialects that cannot be interpreted are treated as noise.

[0089] By using the voice conversion module (140) in this way, the second data and the user's personal information are analyzed to provide a voice tendency that is customized to each user and familiar, thereby reducing resistance to the dementia test diagnosis process using artificial intelligence call and creating a comfortable environment, which can increase the completion rate of the dementia test.

[0090] The memory (150) stores at least one process for performing operations and stores the user's information input and data. A large amount of data can be stored in the memory, and a database system can be installed. In order for the dementia test process to proceed smoothly, personal information of users who have previously undergone dementia testing can be stored, and a database of people in a specific age group (e.g., under 65 years of age) can be stored and managed. This database of test subjects is transmitted to the processor (130), and in order for the dementia test device (100) to operate efficiently, it can provide criteria for controlling that only people of a certain age or older can undergo dementia testing. In addition, the test results and test history of the people included in the database of test subjects can be continuously updated, stored, and managed. By excluding those with a recent test history and selectively sending AI calls only to test subjects who actually need testing, meaningless AI calls can be prevented. The dementia test device (100) manages a database of test subjects loaded in memory (150) and may initiate an AI call-based dementia test outgoing call only for those who have no test history within a preset period.

[0091] The communication module (160) performs the role of enabling the dementia test device (100) to communicate with an external device (200). It communicates with the external device (200) to provide the dementia test process, transmits a question voice to the external device (200), receives an answer voice from the external device (200), and can transmit the results of the dementia test process. It may include at least one of a short-range communication module, a wired communication module, and a wireless communication module, and voice communication and data communication are possible simultaneously. For example, the short-range communication module may include various short-range communication modules that transmit and receive signals using a wireless communication network at a short range, such as a Bluetooth module, an infrared communication module, an RFID communication module, a WLAN communication module, an NFC communication module, and a Zigbee communication module. The wired communication module may include various wired communication modules such as a local area (LAN) module, a wide area (WAN) module, or a value-added communication (VVAN) module, as well as various cable communication modules such as USB, HDMI, DVI, RS-232, power line communication, or POTS. The wireless communication module may include a wireless communication module that supports various wireless communication methods such as Wi-Fi, GSM, CDMA, WCDMA, UMTS, TDMA, LTE, and 5G.

[0092] The STT module (170) refers to a module that converts speech into text. This is explained with reference to FIG. 7. The STT module (170) may include various STT engines, such as, for example, the Google Speech-to-Text API, the Microsoft Azure Speech API, IBM Watson Speech to Text, and Naver™’s Crova Engine. Each of these STT models operates based on unique algorithms and training data for converting speech into text, and can provide different results. By configuring multiple STT models, text conversion can be performed with higher accuracy even in cases where the speech contains specific pronunciations, intonations, or noise. The STT module (170) applies three or more STT models and compares the text results obtained from each model.

[0093] In this process, the text words generated by each STT model are listed to identify identical or similar results, and words or phrases that make up the majority of the converted text are selected. For example, if two or more models produce the same word in the result of converting the same answer voice, that result is considered accurate. The text selected by majority vote is considered a highly reliable result and is subsequently reflected in the dementia test process. The final text conversion result selected by majority vote is used as part of the dementia test process, and the converted text serves as the output result of the STT module (170) and forms the basis for test scoring. The STT module (170) analyzes the content of the user's answer voice through accurate text conversion and evaluates and scores each answer voice. This scored result is used as a basis for judgment in the dementia test process and serves as an important factor in evaluating the user's cognitive ability, language memory, expressive ability, etc.

[0094] FIG. 2 is an exemplary diagram regarding the process of a dementia test being performed by a dementia test device according to one embodiment of the present disclosure.

[0095] As illustrated in FIG. 2, a dementia test according to one embodiment of the present disclosure can be performed by a process of performing a question-and-answer test for dementia with respect to an external device (200) connected to an artificial intelligence call, and converting the answer voice obtained from the user's external device (200) into text data and analyzing it.

[0096] First, the dementia test device (100, see FIG. 1) that receives an inbound request containing first data including a dementia test request from an external device (200) sends an artificial intelligence call to the external device (200) of the test subject in accordance with the inbound request. The dementia test device (100) performs a voice question-and-answer-based dementia test to the user external device (200) connected to the artificial intelligence call, and during the test, the test subject's answer voice can be acquired through the external device (200) and transmitted to the dementia test device (100).

[0097] The dementia test device (100) can convert the answer voice of a test subject received from the user's external device (200) into text data by the STT module (170). Meanwhile, the answer voice of the test subject may include not only words intended to answer questions provided by the test subject during the dementia test, but also other words not intended to answer questions, such as interjections and self-talk. For example, referring to the text data converted from the answer voice shown in FIG. 2, the answer voice may include "dog," "cat," "what was there," "horse," etc. Here, "dog," "cat," and "horse" correspond to words intended to answer questions provided by the test subject during the dementia test and are content that can be meaningfully utilized in the analysis of the test content. On the other hand, "what was there" belongs to self-talk not intended to answer questions and is therefore content that does not help in the analysis of the test content. Words belonging to self-talk are treated as noise and are not included in the process of calculating the dementia test score. The dementia test device (100) can improve the efficiency and accuracy of the test by extracting only the words that can be meaningfully utilized in the analysis of the test content from the text data and performing the analysis.

[0098] The dementia testing device (100) calculates a quantified value that can determine whether a test subject has dementia or the level of dementia symptoms based on extracted words. By comparing and analyzing the quantified value with a pre-stored database of test groups, the dementia testing device (100) can determine whether the test subject needs to visit a testing institution for follow-up testing.

[0099] FIG. 3 is a conceptual diagram relating to a part of an inbound request of a dementia test device according to one embodiment of the present disclosure.

[0100] The dementia test process of the inbound dementia test device (100) is initiated by first data including a dementia test request. The first data can be generated by various types of inbound requests (300). For example, it can be collected through one or more methods such as telephone calls, scanning of a two-dimensional barcode, inputting a web address, sending SMS text messages, app push touches, NFC tag contacts, and kiosk inputs. The first data is mostly generated by an external device (200), such as a smartphone owned by a user who wishes to participate in the dementia test, but it may also be generated by a desktop PC, kiosk, or computer connected to an input device, rather than a smartphone, and transmitted to the dementia test device (100). The external device (200) may be a mobile device owned by the test subject, or a tablet PC, desktop PC, kiosk, or computer connected to an input device.

[0101] Since the first data is generated by a user possessing an external device (200), if it is not properly controlled, the server may crash or the network may be paralyzed due to the occurrence of an excessive dementia test process. If excessive inbound requests occur, it leads to a waste of budget and can have a significant impact on the progress of the dementia management business. Therefore, a step to verify whether the subject is a test subject (S411) is absolutely necessary, and the inbound dementia test device must handle the 'dementia test request' precisely. In order to prevent the occurrence of repetitive dementia test request signals, it is necessary to analyze information regarding the external device (200) from which the first data was generated and transmitted, and to determine whether it is the same inbound request (300) from the same user.

[0102] Since the inbound request (300) can be transmitted to the dementia testing device (100) by various methods such as phone call, two-dimensional barcode scanning, web address input, SMS text transmission, app push touch, NFC tag contact, kiosk input, etc., it must be determined whether the request is made by the same user, and in order to determine this, second data and personal information consent may be required. The inbound request (300) is generated by an external device (200), and since the external device (200) is a device connected to communication such as a smartphone, computer, kiosk, or tablet PC, it can analyze the unique protocol for communication and obtain reference information to determine whether it is the same user based on a unique address such as a MAC address.

[0103] FIG. 4 is a flowchart schematically illustrating a part of an inbound request-based dementia test method according to one embodiment of the present disclosure.

[0104] As illustrated in FIG. 4, a dementia test method according to one embodiment of the present disclosure may include an inbound request step (S410), a step of verifying whether the subject is a test subject (S411), a step of providing guidance voice (S420), a first test progress step (S451, S453), a second test progress step (S471, 473), a test content analysis step (S481), and a test result transmission step (S491).

[0105] The inbound request step (S410) is a step for collecting first data containing a user's dementia test request. The dementia test process of the inbound dementia test device (100) is initiated by the first data containing the dementia test request. The first data can be collected through one or more methods including telephone outgoing, two-dimensional barcode scanning, web address input, SMS text transmission, app push touch, NFC tag contact, and kiosk input. The first data is mostly generated by an external device (200), such as a smartphone owned by a user who wishes to participate in the dementia test, but it may also be generated by a desktop PC, kiosk, or computer connected to an input device, rather than a smartphone, and transmitted to the dementia test device (100). The external device (200) may be a mobile device owned by the test subject, or a tablet PC, desktop PC, kiosk, or computer connected to an input device. From the perspective of the dementia test device (100), the 'dementia test request' is received first, so it is called an 'inbound request'.

[0106] The step of verifying whether the subject is a test subject (S411) is a step to prevent repeated requests for dementia tests by the same user. Although it is possible for the same user to verify whether dementia is improving by obtaining dementia test results at regular intervals using the dementia test device (100) according to one embodiment of the present disclosure, it is necessary to prevent repeated tests by the same person in advance because the budget for dementia tests conducted by dementia management institutions and medical institutions is limited. In addition, since there is a need to perform dementia tests on potential dementia patients of a specific age group or older by administrative district, the step of verifying whether the subject is a test subject (S411) allows only users of a certain age or older to proceed with the dementia test process.

[0107] The guidance voice provision step (S420) is a step in which the dementia test device (100) provides guidance voice to a user who is a subject of the dementia test through an external device (200). In one embodiment, the guidance voice may include information on the subject (recipient), information on the testing institution, information on the test solution (program), and information on the reward provided upon completion of the test in the first part. At this time, the information on the subject (recipient) is identified based on the first data and second data identified in the aforementioned inbound request step, and by being inserted into the guidance voice, the guidance voice can be provided in a customized manner for each subject. Through this, the test participation rate and test completion rate of the subject who received the AI ​​call can be improved. After the guidance voice is provided, the dementia test may proceed, and the dementia test may consist of a first test and a second test that proceed sequentially.

[0108] The first test progress step (S451, S453) is a step for conducting a first test, which is a practice test, and improving the test subject's (recipient of the external device (200)'s understanding of the test progress method. The first test progress step may include a step (S451) in which the dementia test device (100) provides a first question voice to the external device (200), and a step (S453) in which the dementia test device (100) obtains a first answer voice from the external device (200).

[0109] Here, the first question voice may include a voice requesting an answer to the first topic during the first time, and the first answer voice may include a voice in which the test subject answers in response to the first question voice.

[0110] The subject of the test may participate in the first test by responding to the voice of the first question, and the subject's first answer voice may be acquired through the sound acquisition unit of the external device (200) and transmitted to the dementia test device (100) through the communication unit of the external device (200).

[0111] In one embodiment, a first response voice obtained through a first test can be used as data to determine whether the test subject has properly understood the test procedure. That is, the artificial intelligence call-based dementia test method according to one embodiment of the present disclosure may further include a step of determining the test subject's understanding of the test procedure based on the first response voice.

[0112] For example, the dementia test device (100) can calculate a value of understanding of the test subject's test procedure and determine whether to proceed with the test based thereon. Here, "understanding value" is a numerical value representing the degree of understanding of the test subject's test procedure. As a specific example, the dementia test device (100) calculates a value of understanding based on the ratio of the word voice that answered in response to the first question voice and other voices among the first answer voices, and if the calculated value of understanding is less than a preset value, the second test may not be conducted.

[0113] As another specific example, the dementia test device (100) calculates a comprehension value based on the number of words the test subject answered regarding the first topic presented in the first test, and if the calculated comprehension value is less than a preset value, the second test may not be conducted. If the second test is not conducted because it is determined that the test subject did not properly understand the test procedure, the dementia test device (100) may provide a voice guiding the test procedure or schedule a subsequent test, but is not limited to this.

[0114] In this case, the dementia test device (100) may provide a voice guiding the test procedure or schedule a test date thereafter, but is not limited thereto.

[0115] In another embodiment, the step of obtaining a first answer voice from an external device (200) during the first test process step may be omitted. Since the first test is merely a test intended to improve understanding of the test process method and is not used as data to judge the subject's language fluency, obtaining the answer voice for the first test may be omitted. This allows for a reduction in the total amount of data transmitted and received during the test process.

[0116] The second test procedure step (S471, 473) is a step of obtaining voice data used to determine the presence or absence of dementia and the level of dementia symptoms while conducting the second test, which is the main test.

[0117] Referring to FIG. 4, the second test progress step may include the step (S471) of the dementia test device (100) providing a second question voice to an external device (200) and the step (S473) of the dementia test device (100) obtaining a second answer voice from the external device (200).

[0118] Here, the second question voice may include a voice requesting an answer to the second topic during the second time, and the second answer voice may include a voice in which the test subject answers in response to the second question voice.

[0119] The subject of the examination may participate in the second examination by responding to the second question voice, and the subject's second answer voice may be acquired through the sound acquisition unit of the external device (200) and transmitted to the dementia examination device (100) through the communication unit of the external device (200).

[0120] The test content analysis step (S481) is a step in which the dementia test device (100) analyzes the test content based on the second answer voice obtained from the external device (200).

[0121] In the test content analysis step (S481), the dementia test module (120, see FIG. 1) of the dementia test device (100) can convert the second answer voice received from the external device (200) into text data and calculate a language fluency value based on the converted text data to analyze the test content. Specifically, the test content analysis step (S481) may include the step of converting the second answer voice received from the external device (200) into text data; the step of extracting at least one word corresponding to the second topic of the second test from the converted text data; the step of calculating a language fluency value based on the at least one word corresponding to the extracted second topic; the step of comparing and analyzing the language fluency value with a preset reference value; and the step of determining the subject as a follow-up test subject if the language fluency value is smaller than the preset reference value, and determining the subject as a non-subject for the follow-up test if the language fluency value is greater than or equal to the preset reference value.

[0122] The step of calculating the language fluency value may include a step of scoring by applying addition or subtraction criteria to at least one of the total number of words, the number of words in the first half, the number of words in the second half, the number of characters per word, the number of category changes, the number of words per category, and the number of duplicate words. Additionally, the preset threshold value for comparing the calculated language fluency value may be differentially set for each of the test groups classified based on auxiliary information including at least one of gender, age, education level, and number of cohabitants. The calculated language fluency value may be compared and analyzed with the threshold value set for the group to which the test subject belongs.

[0123] Meanwhile, in one embodiment, the test content analysis step (S481) may further include a step of analyzing changes in speech style based on the existing test data of the test subject and the second answer voice obtained through this test, in the case where the test subject has a test history, that is, when the external device (200) connected to the artificial intelligence call has a test history. Here, "speech style" may include speech speed, pronunciation accuracy, etc., but is not limited thereto.

[0124] The dementia test module (120, see FIG. 1) of the dementia test device (100) can analyze the speech style of the second answer voice obtained in this test and calculate a speech style change value by comparing it with data on the speech style of the answer voice obtained in the previous test. The dementia test module (120) of the dementia test device (100) can determine a subject for a follow-up test if the speech style change value is greater than or equal to a specific value (a preset value). In addition, the dementia test module (120) of the dementia test device (100) can determine a subject for a follow-up test if the difference between the language fluency value calculated in the previous test and the language fluency value calculated in this test is greater than or equal to a specific value (a preset value). Through this, not only the absolute language fluency value but also cases where language fluency has relatively decreased compared to the time of the previous test can be tested to perform a follow-up test, thereby having the effect of preventing the rapid deterioration of symptoms in advance.

[0125] In one embodiment, the test content analysis step (S481) may further include a step of determining the level of understanding of the test procedure based on the first answer voice. Specifically, the dementia test device (100) can analyze the first answer voice to determine whether the test subject fully understands the test procedure and is answering, or whether the test subject does not understand and is giving answers unrelated to the test content. For example, the dementia test device (100) may determine that the test subject has properly understood the test procedure if the test subject answers more than a preset number of words regarding the first topic presented in the first test, and determine that the test subject has not properly understood the test procedure if they do not.

[0126] Even if the language fluency value calculated by analyzing the second answer voice is below a preset threshold value, if the understanding value of the test procedure calculated by analyzing the first answer voice is below a specific value (preset value), the dementia test device (100) may not immediately determine the subject as a subject for a subsequent test, but may proceed with a re-test after providing re-guidance on the test procedure. Through this, it is possible to distinguish cases where a test subject with excellent language fluency (no symptoms of dementia) receives a low language fluency value due to a lack of understanding of the test procedure, thereby improving the accuracy of the test.

[0127] The inspection result transmission step (S491) is a step of transmitting the inspection result analyzed in the inspection content analysis step (S481) to an external device (200).

[0128] In one embodiment, the test results may be transmitted in the form of a message, and the message may include a link to check the test results. The method of transmitting the test results is not limited to the examples described above and may be provided in various forms, such as telephone, mail, or messages. Furthermore, the recipient of the test results is not limited to the test subject but may include various people in the subject's vicinity, such as cohabitants or guardians.

[0129] In one embodiment, the test results may include information on whether the subject corresponds to a subsequent test subject and information on predicting subsequent symptoms. Here, the information on predicting subsequent symptoms is derived based on symptom information of a similar group of people having auxiliary information similar to the test subject (e.g., age, gender, education level, number of cohabitants, etc.), and may include the trend of decrease in language fluency values ​​over time, results of subsequent tests, etc.

[0130] FIG. 5 is an example diagram regarding the addition of additional content using location information of an external device in a dementia test process according to one embodiment of the present disclosure. Based on FIG. 5, an inbound request-based dementia test device process linked with location information of an external device (200) will be described.

[0131] The processor (130) can identify the location information at the time of transmission of the external device (200) from which the first data was transmitted, and based on the identified location information, generate first jurisdiction information regarding the administrative district where the external device (200) is located and second jurisdiction information regarding the medical institution responsible for the location. The processor (130) of the dementia test device (100) can generate medical institution guidance or additional tests based on the first jurisdiction information and the second jurisdiction information to be additionally configured in the basic dementia test.

[0132] The process of ‘generating jurisdictional information based on location information at the time of transmission by the external device’ is explained using Gyeonggi-do as an example. The process of identifying location information is as follows. The dementia test device (100) identifies location information transmitted by the external device (200). For example, assuming that potential dementia patient A, who resides in Gyeonggi-do, sends a request for a dementia test via their smartphone, the location at the time of transmission is confirmed to be ‘Paldal-gu, Suwon-si, Gyeonggi-do’ through the smartphone's GPS data. Subsequently, administrative district jurisdictional information is generated. Based on the identified location information, the dementia test device extracts administrative district information such as ‘Gyeonggi-do’ and ‘Paldal-gu, Suwon-si’. Through this process, the first jurisdictional information, that is, information regarding the administrative district to which A belongs, is generated. This plays an important role as the first step in personalizing the dementia test process. Subsequently or simultaneously, medical institution jurisdictional information is generated. Based on the extracted administrative district information, the dementia test device verifies information about the medical institution responsible for the area. For example, in Paldal-gu, Suwon-si, Gyeonggi-do, the ‘Suwon Dementia Relief Center’ may be a medical institution responsible for dementia-related diagnosis and management. Therefore, the dementia testing device generates information called 'Suwon Dementia Care Center' as second jurisdictional information.

[0133] The generated first jurisdiction information (Gyeonggi-do and Paldal-gu, Suwon-si) and second jurisdiction information (Suwon Dementia Relief Center) are utilized in the dementia testing process in the following ways. The dementia testing device (100) can transmit the test results of a user who has taken the dementia test to the Suwon Dementia Relief Center so that the relevant institution can provide follow-up measures. In addition, it can provide information on medical institutions near the user who has taken the dementia test to encourage a direct visit if necessary. Precise test results can be derived by configuring a customized set of dementia-related questions that reflect the characteristics of the Gyeonggi-do region.

[0134] This process provides practical assistance to potential dementia patients and their guardians residing in Gyeonggi-do, creating an environment where they can receive examinations and medical services specialized for their residential area. By connecting them to appropriate medical institutions based on the test results, time and costs are saved, and dementia prevention and management services can be provided quickly and efficiently. In this way, the dementia testing device (100) can generate administrative district and medical institution information based on the user's location information, thereby optimizing the testing process and follow-up measures to suit the user.

[0135] FIG. 6a is a conceptual diagram showing the time-series process of a dementia test process according to one embodiment of the present disclosure, and FIG. 6b is a conceptual diagram showing the time-series process of a dementia test process with additional content added using location information of an external device according to one embodiment of the present disclosure.

[0136] The processor (130) subdivides the second jurisdiction information into metropolitan medical institutions and local medical institutions, and according to the second jurisdiction information, can add guidance for each medical institution and additional examination processes to the voice question-and-answer-based dementia examination process provided to the external device (200) by default. The processor (130) subdivides the first jurisdiction information into metropolitan administrative institutions and local administrative institutions, and analyzes the number of dementia examination requests, request methods, request locations, and request times of the user based on the first data, second data, and first jurisdiction information, and can process the dementia examination process so as not to proceed if the same user requests more than a preset number of dementia examinations within a preset period.

[0137] FIG. 6a illustrates the chronological process of a dementia test process according to one embodiment of the present disclosure. A voice-based basic dementia test process consists of an introduction section, a first test section, a second test section, and a conclusion section. The dementia test process can be performed by including a dementia test module (120) that, after receiving first data from an input module (110), establishes a telephone call with the external device (200) for the dementia test, provides guidance voice through a communication module (160), conducts a first test which is a preliminary practice test to improve the subject's understanding of the test procedure, and conducts a second test which is a main test used to determine the presence or absence of dementia disease or the level of dementia symptoms of the subject.

[0138] The first test may be conducted by providing the external device (200) with a first question voice requesting an answer to a first topic for a first time and obtaining a first answer voice from the external device (200), and the second test may be conducted by providing the external device (200) with a second question voice requesting an answer to a second topic different from the first topic for a second time longer than the first time and obtaining a second answer voice from the external device (200). The first test is intended to have the effect of pre-practicing the user taking the dementia test, and the actual dementia test is not performed. The actual dementia test is performed through the second test. Therefore, the time required for the second test is longer, and the first and second tests combined do not exceed 3 minutes.

[0139] FIG. 6b is a conceptual diagram showing the time-series process of a dementia test process with additional content added using location information of an external device according to one embodiment of the present disclosure. When additional content using location information of an external device is added according to one embodiment of the present disclosure, medical institution guidance (t0) may be added between the introduction part and the first test part, and additional test (t3) may be added between the second test part and the conclusion part. This is performed by a processor (130), and the processor (130) can receive voice information regarding medical institution guidance, additional test, etc. from each administrative agency and medical institution based on first jurisdiction information, first data, and second data, and perform processing to automatically edit or insert it into the basic voice test. The order of medical institution guidance and additional test is not limited to the order shown in FIG. 6b, and the content may also be changed.

[0140] FIG. 7 is an exemplary diagram relating to an STT multiplexing process in a dementia testing device according to one embodiment of the present disclosure.

[0141] The processor (130) controls an STT module (170) composed of one or more Speech-to-Text (STT) models, applies three or more STT models to convert the answer voice into text, compares the converted text words, and selects the converted text that accounts for the majority of the converted texts to use for scoring the results of the dementia test process. This is a type of voting system for the accuracy of the STT process. Since elderly dementia test participants may have unclear pronunciation, accurate voice analysis may be difficult with existing STT engines. Therefore, it is desirable to use three or more STT engines to convert speech to text using the closest word. For example, a dementia test participant may pronounce "tiger," but a specific STT engine may convert it into text as "gorani." However, if the other two STT engines convert it into text as "tiger," the word "tiger" can be selected in a 1:2 ratio to proceed with the dementia test process. This is a process that increases the accuracy of dementia testing.

[0142] FIG. 8 is a flowchart schematically illustrating a part of a dementia test method according to one embodiment of the present disclosure.

[0143] In an inbound request-based dementia test method performed by a device, the method comprises the steps of: collecting first data containing a user's request for a dementia test by one or more methods including telephone calling, two-dimensional barcode scanning, web address input, SMS text transmission, app push touch, NFC tag contact, and kiosk input; collecting second data containing one or more of the user's name, phone number, gender, and highest level of education; and collecting consent information for the use of personal information regarding the use of said second data; identifying location information at the time of transmission of said external device (200) from which the first data was transmitted; generating first jurisdiction information regarding the administrative district where said external device (200) is located and second jurisdiction information regarding the medical institution responsible for said location based on the identified location information; subdividing said second jurisdiction information into metropolitan medical institutions and basic medical institutions; and, according to said second jurisdiction information, adding guidance for each medical institution and an additional test process to said voice question-and-answer-based dementia test process that is basically provided to said external device (200). An inbound request-based dementia test method is disclosed, comprising the steps of transmitting a question voice to the external device (200), receiving an answer voice from the external device (200), calculating a score by the dementia test process, and transmitting the calculated score and the result to the user's external device (200).

[0144] The steps of the method or algorithm described in connection with the embodiments of the present disclosure may be implemented directly in hardware, implemented as a software module executed by hardware, or implemented by a combination thereof. The software module may reside in RAM (Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), Flash Memory, a hard disk, a removable disk, a CD-ROM, or any form of computer-readable recording medium well known in the art to which the present disclosure belongs.

[0145] Although embodiments of the present disclosure have been described above with reference to the attached drawings, those skilled in the art will understand that the present disclosure may be implemented in other specific forms without altering its technical concept or essential features. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive.

[0146] [Explanation of the symbol]

[0147] 100: Dementia testing device

[0148] 110: Input module

[0149] 120: Dementia Screening Module

[0150] 130: Processor

[0151] 140: Speech conversion module

[0152] 150: Memory

[0153] 160: Communication module

[0154] 170: STT Module

[0155] 200: External device

[0156] 300: Inbound request

Claims

1. A dementia testing device that provides a non-face-to-face voice question-and-answer-based dementia testing process, An input module that collects first data including a user's request for a dementia test through at least one method among telephone calling, code scanning, web address input, text message transmission, app push touch, NFC tag contact, and kiosk input; A communication module that communicates with an external device to provide the dementia test process, transmits a question voice to the external device, receives an answer voice from the external device, and transmits a result based on the dementia test process; A memory in which the above voice question-and-answer-based dementia test process is stored; and It includes a processor that performs an operation according to the above process, The above processor is, Collecting second data including at least one of the user's name, phone number, gender, and highest level of education, and personal information consent information regarding the use of the second data together with the first data through the input module, and Analyzes information regarding the external device to which the above first data was generated and transmitted, and Based on information regarding the above external device, first data, and second data, a dementia test process tailored to the user's environment is configured, and Identifying the location information of the external device at the time of transmission from which the first data was transmitted, and generating first jurisdiction information regarding the administrative district where the external device is located and second jurisdiction information regarding the medical institution responsible for the location based on the identified location information, The above second jurisdictional information is subdivided into metropolitan medical institutions and local medical institutions, and according to the above second jurisdictional information, guidance for each medical institution and additional testing processes are added to the voice question-and-answer-based dementia testing process that is basically provided to the above external device, and Based on the above first jurisdictional information, subdivide into metropolitan administrative agencies and local administrative agencies; based on the above first data, the above second data, and the above first jurisdictional information, analyze the number of dementia test requests, the request method, the request location, and the request time of the user; and in the case of dementia test requests exceeding a preset number within a period by the same user, process so as not to proceed with the above dementia test process. The received response voice is analyzed in real time to separate it into the user's voice and background voice, and if the background voice includes the voice of a person other than the user or if it is determined that the background voice was generated in a place where noise of 50 decibels or more exists, a decision on whether to stop the dementia test process is transmitted to the external device. An inbound request-based dementia testing device that analyzes the word count, first-half word count, second-half word count, character count per word, number of category changes, word count per category, speech speed, pronunciation accuracy, intonation, stress change, voice tremor, topic consistency, maintenance of conversational flow, and number of duplicate words based on the received response voice, and scores the results of the dementia testing process by applying criteria for adding or subtracting points to the analyzed results.

2. In Paragraph 1, The above processor is, When the first data is received from the input module, a telephone call is established with the external device for the dementia test, and Providing guidance voice through the above communication module, and conducting a first test, which is a preliminary practice test, to enhance the test subject's understanding of the test procedure, and An inbound request-based dementia testing device that conducts a second test, which is the main test used to determine the presence or absence of dementia disease or the level of dementia symptoms of the above-mentioned test subject.

3. In Paragraph 2, The above first test is an inbound request-based dementia test device that proceeds by providing a first question voice requesting an answer to a first topic for a first time period to the external device and obtaining a first answer voice from the external device.

4. In Paragraph 2, An inbound request-based dementia test device, wherein the second test is performed by providing the external device with a second question voice requesting an answer to a second topic different from the first topic for a second time longer than the first time, and obtaining a second answer voice from the external device.

5. In Paragraph 1, The above processor is, Controls an STT module composed of one or more Speech-to-Text (STT) models, and Applying three or more speech-to-text conversion models to convert the above-mentioned answer speech into text, comparing the words of the converted text, and selecting the text that accounts for the majority of the converted text, An inbound request-based dementia testing device that uses the selected text above to score the results of the dementia testing process above.

6. A method for providing a non-face-to-face voice question-and-answer-based dementia screening process performed by a device, A step of collecting first data including a user's request for a dementia test by at least one method among phone call, code scanning, web address input, text message transmission, app push touch, NFC tag contact, and kiosk input, second data including at least one of the user's name, phone number, gender, and highest level of education, and consent information for the use of personal information regarding the use of said second data; A step of identifying the location information of the external device at the time of transmission of the first data, generating first jurisdiction information regarding the administrative district where the external device is located and second jurisdiction information regarding the medical institution responsible for the location based on the identified location information, subdividing the second jurisdiction information into metropolitan medical institutions and local medical institutions, and adding guidance for each medical institution and an additional examination process to the voice question-and-answer-based dementia examination process basically provided to the external device according to the second jurisdiction information; A step of identifying the location information of the external device at the time of transmission of the first data, and generating first jurisdiction information regarding the administrative district where the external device is located and second jurisdiction information regarding the medical institution responsible for the location based on the identified location information; A step of subdividing the above second jurisdictional information into metropolitan medical institutions and local medical institutions, and, according to the above second jurisdictional information, adding guidance for each medical institution and additional testing processes to the voice question-and-answer-based dementia testing process basically provided to the above external device; A step of subdividing into metropolitan administrative agencies and local administrative agencies based on the first jurisdictional information, analyzing the number of dementia test requests, request method, request location, and request time of the user based on the first data, the second data, and the first jurisdictional information, and processing so as not to proceed with the dementia test process in the case where the same user requests more than a preset number of dementia tests within a period; A step of analyzing the received response voice in real time to separate it into the user's voice and background voice, and transmitting to the external device whether to stop the dementia test process if the background voice includes the voice of a person other than the user or if it is determined that the background voice was generated in a place where noise of 50 decibels or more exists; A step of analyzing the word count, word count in the first half, word count in the second half, character count per word, number of category changes, word count per category, speech rate, pronunciation accuracy, intonation, stress change, voice tremor, topic consistency, maintenance of conversational flow, and number of duplicate words based on the received response voice, and scoring the results of the dementia test process by applying criteria for adding or subtracting points to the analyzed results: A method comprising the steps of transmitting a question voice to the external device, receiving an answer voice from the external device, calculating a score based on the dementia test process, and transmitting the calculated score and the result to the external device.

7. In Paragraph 6, The above device is, When the first data is received from the input module, a telephone call is established with the external device for the dementia test, and Providing guidance voice through the above communication module, and conducting a first test, which is a preliminary practice test, to enhance the test subject's understanding of the test procedure, and A method for conducting a second test, which is the main test used to determine the presence or absence of dementia or the level of dementia symptoms of the subject of the above test.

8. In Paragraph 7, The above first test is a method of conducting by providing a first question voice requesting an answer to a first topic for a first time to the external device, and obtaining a first answer voice from the external device.

9. In Paragraph 7, The above second test is a method of providing the external device with a second question voice requesting an answer to a second topic different from the first topic for a second time longer than the first time, and obtaining a second answer voice from the external device.

10. In Paragraph 6, The above processor is, Controls an STT module composed of one or more Speech-to-Text (STT) models, and Applying three or more speech-to-text conversion models to convert the above-mentioned answer speech into text, comparing the words of the converted text, and selecting the text that accounts for the majority of the converted text, A method for using the selected text above to score the results of the dementia test process above.

Images