system

Abstract

We provide the system. [Solution] A means for capturing and pre-processing viewer activity data in real time, A means of extracting behavioral pattern features from captured data, A method that utilizes machine learning models to predict responses to inquiries and applications based on extracted features, A means of generating and proposing strategic reports on marketing activities based on prediction results, A system that includes this.

Description

Technical Field

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

Background Art

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

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In recent years, in product explanation meetings that utilize video content, it has been required to effectively capture the actions of viewers and predict the possibility of inquiries and applications based on such data. However, it is difficult to conduct precise analysis by utilizing real-time action data of viewers and promptly formulate a marketing strategy based on such analysis. There is a need for a method to improve such a situation and realize efficient sales activities through video content.

Means for Solving the Problems

[0005] This invention provides a system for collecting and preprocessing viewer activity data in real time. By extracting behavioral pattern features from the collected data and applying a machine learning model that predicts responses to inquiries and applications using these features, the system generates and proposes strategic reports for marketing activities based on the prediction results. Furthermore, this system generates identification information using event data such as viewing time and playback, which indicate viewer interest, and provides feedback useful for optimizing video content by providing reports optimized for remote terminals.

[0006] "Viewer activity data" refers to data recorded when viewers watch product presentation videos, including playback, pausing, skipping, and viewing time.

[0007] "Real-time capture" means instantly detecting and recording a series of actions that viewers take while watching a video.

[0008] "Preprocessing" refers to preparing collected raw data into an analyzable format using techniques such as organizing, interpolating missing values, and removing outliers.

[0009] "Behavioral pattern features" are indicators extracted based on viewers' viewing behavior, and they quantify interests and concerns, such as average viewing time and viewing completion rate.

[0010] A "machine learning model" is a collection of algorithms and computational methods that learn from past data and use that learning to make predictions and classifications on new data.

[0011] "Inquiry and application responses" refer to the actions viewers take regarding products or services after watching a video, such as making inquiries or applications.

[0012] A "strategic marketing activity report" is a report generated based on forecast results, providing marketing professionals with guidance for their next course of action.

[0013] "Identification information" refers to data generated based on viewer behavior data to indicate individual behavioral characteristics and levels of interest.

[0014] "Feedback" refers to advice and suggestions for improvement provided based on the analysis results, and is information used to optimize video content. [Brief explanation of the drawing]

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

Mode for Carrying Out the Invention

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

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

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

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

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

[0021] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0022] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0023] [First Embodiment]

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

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

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

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

[0028] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0029] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0030] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

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

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

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

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

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

[0036] This invention is a system for analyzing viewer activity data and optimizing marketing activities based on that data. This system is broadly composed of the following processes: data collection, data analysis, predictive model generation, and result feedback.

[0037] First, the server collects diverse data in real time from viewers of the product presentation video. It captures behavioral data such as viewing time, play / pause events, and skipped sections, and converts it into a format that can be immediately analyzed. This process prepares the foundational data necessary to understand the behavioral characteristics of the viewers.

[0038] Next, the server performs analysis based on the collected data. It extracts viewer behavior patterns as features and uses them to run a machine learning model to predict the likelihood of inquiries or applications. This makes it possible to numerically visualize how interested viewers are in the product. This model is designed to make highly accurate predictions by utilizing past training data.

[0039] The server then analyzes the prediction results and generates a strategic report in a format that the marketing team can use. This report includes content that resonated with viewers, responses by segment, and areas for improvement, and is provided to each user's device. This makes it easier for the marketing team to make decisions regarding adjusting video content and executing their next campaign strategies.

[0040] For example, when a user watches a video, sections where they frequently rewind or watch for extended periods are recorded. The server analyzes this data to determine which sections strongly capture the viewer's interest. This information is reflected in reports and used by the marketing team as valuable insights to create new promotional videos that highlight those sections.

[0041] The following describes the processing flow.

[0042] Step 1:

[0043] The server monitors the activity of users watching the product presentation videos in real time and collects data from viewers. Specifically, it acquires and records detailed event data such as the start time of viewing, the timing of playback and stopping, the sections skipped, and the viewing completion rate.

[0044] Step 2:

[0045] The server performs preprocessing on the collected raw data. Preprocessing includes organizing the data, interpolating missing values, and removing noisy data. This process creates a well-organized dataset that is suitable for subsequent analysis.

[0046] Step 3:

[0047] The server extracts viewer behavior patterns as features from the pre-processed data. In this step, multidimensional features such as average viewing time, rewind frequency, and skipping tendencies are calculated. This quantifies viewer interest and engagement.

[0048] Step 4:

[0049] The server uses a machine learning model based on extracted features to predict the likelihood of inquiries or applications. The machine learning model is pre-trained to achieve precise predictions. The model's output represents the probability of a response corresponding to each viewing session.

[0050] Step 5:

[0051] The server receives the prediction results and generates a visual strategic report. This report includes interesting points about viewing patterns and areas for improvement, providing concrete suggestions for marketing activities.

[0052] Step 6:

[0053] The terminal receives strategic reports provided by the server, which are then reviewed by members of the marketing team. This helps them optimize their video content editing and distribution strategies for the future.

[0054] (Example 1)

[0055] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0056] When building effective marketing strategies using viewer behavior data, it is essential to accurately understand viewer interests and efficiently optimize strategies. However, currently, the process from data collection to analysis and strategy proposal is fragmented, making it difficult to respond quickly and make highly accurate predictions.

[0057] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0058] In this invention, the server includes means for collecting and preprocessing viewer behavior information from terminals in real time, means for extracting feature quantities that indicate behavioral trends from the collected behavior information, and means for utilizing a generative AI model that predicts responses to inquiries and suggestions based on the extracted feature quantities. This enables a rapid and highly accurate understanding of viewer interests and the optimization of effective marketing strategies.

[0059] "Device" refers to an electronic device used by viewers to watch video, and includes personal computers, smartphones, tablets, and other similar devices.

[0060] "Viewers" refers to individuals or groups who watch video content and whose viewing behavior is collected as data.

[0061] "Action information" refers to data about the actions viewers perform while watching a video, and specifically includes viewing time, playback, pause, skip, and other events.

[0062] "Collecting data in real time" refers to the process of instantly sending viewer behavior information to the server without delay after an action occurs and acquiring it as data.

[0063] "Preprocessing" refers to the data manipulation required to prepare collected raw data for analysis, and includes imputing missing values, removing outliers, and converting to an appropriate data format.

[0064] In data analysis, "features" refer to data that quantifies patterns or statistical properties used by models for learning and prediction.

[0065] A "generative AI model" refers to an artificial intelligence model that is built to extract regularities and patterns from given data and to make predictions and classifications based on new data.

[0066] "Market research activities" refer to a series of analyses and strategy-building processes conducted to understand market trends for a particular product or service.

[0067] A "strategic report" is a document used to consider the next marketing steps based on the results of market research activities, and refers to a report that includes data analysis results and recommended measures.

[0068] This invention relates to a data analysis system for optimizing marketing activities by utilizing viewer behavior data. This system uses a device on which users view video content and collects their viewing behavior as data. The device consists of personal computers, smartphones, tablets, etc. A server collects viewer behavior information from these devices in real time. This behavior information includes viewing time, playback, pause, skip, and other events, and a JavaScript® code snippet runs on the device to send the data to the server.

[0069] Subsequently, the server preprocesses the collected data, imputing missing values ​​and removing outliers, and uses libraries such as Python's Pandas to convert the data into a unified format. From the preprocessed data, the server extracts features of viewer behavior patterns and prepares them as input data for a machine learning model. The generative AI model utilizes a pre-trained neural network using libraries such as Scikit-learn and TENSORFLOW®.

[0070] The server uses a generative AI model to predict viewers' levels of interest, inquiries, and responses to suggestions. This makes it possible to quantify and visualize the degree to which viewers are interested in a product or service. Based on these predictions, the server generates a strategic report for market research activities and provides it to each remote terminal.

[0071] For example, if a user repeatedly plays a specific section while watching a product explanation video, the server will determine that the user has a high level of interest in that section and reflect this information in the report. The marketing team can then use this report to create new promotional content that highlights that section.

[0072] An example of a prompt would be: "Analyze the predictive report generated from viewing data and propose a new promotional strategy that highlights the sections that particularly captured audience interest."

[0073] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0074] Step 1:

[0075] The device collects real-time information about the user's actions while watching videos. This information includes viewing time, playback, pause, skipping, etc. A JavaScript code snippet is used to build a system that periodically sends this data to the server. The input to this process is the user's viewing behavior, and the output is the behavioral data sent to the server.

[0076] Step 2:

[0077] The server preprocesses the operational information received from the terminal. Specifically, it uses the Python Pandas library to impute missing values ​​and filter out outliers. Afterward, it converts the data into a unified format, preparing it for analysis. The input for this step is the raw data received from the terminal, and the output is the preprocessed, purified data.

[0078] Step 3:

[0079] The server extracts features from pre-processed data. It quantifies viewer behavior patterns, such as the distribution of viewing time and the proportion of frequently played sections. Statistical methods and data analysis techniques are used for this feature extraction. The input for this step is pre-processed data, and the output is a feature vector.

[0080] Step 4:

[0081] The server processes feature vectors using a generative AI model to predict viewer interest and likelihood of response. It then applies a pre-trained model using Scikit-learn or TensorFlow to calculate a prediction score. The input for this step is a feature vector, and the output is the prediction score.

[0082] Step 5:

[0083] The server generates a strategic report on market research activities based on the prediction score. Visualization tools are used to format the report in a visually easy-to-understand manner. The report includes content sections that viewers showed particular interest in, as well as recommended improvement actions. The input for this step is the prediction score, and the output is the strategic report.

[0084] Step 6:

[0085] The user (marketing team) uses the generated strategic report to optimize their next marketing strategy. Based on the feedback in the report, they implement specific measures, such as creating new promotional content. The input for this step is the strategic report, and the output is the improved marketing activities.

[0086] (Application Example 1)

[0087] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0088] Traditional methods of utilizing viewer data have made it difficult to efficiently analyze large amounts of data and formulate accurate marketing strategies based on that analysis. Furthermore, providing content that reflects individual viewer interests and behavioral patterns in real time has been challenging, resulting in difficulties in improving customer satisfaction and business performance.

[0089] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0090] In this invention, the server includes means for capturing and pre-processing viewer activity data in real time, means for extracting behavioral pattern features from the captured data, and means for tracking viewer responses to each viewing section in real time and providing personalized content suggestions. This enables direct utilization of viewer reactions to provide dynamic content tailored to customer interests and optimize marketing strategies.

[0091] "Activity data" refers to data about viewers' actions when viewing content, including viewing time and events such as playback, pausing, and skipping.

[0092] "Preprocessing" refers to the process of preparing data into an analyzable format, including tasks such as data cleaning and format conversion.

[0093] "Behavioral pattern features" are analytical indicators extracted from viewer behavior data, and are elements used to quantify viewing trends and interests.

[0094] "Personalized content recommendations" is an approach that presents specific content tailored to an individual's interests based on their behavioral data.

[0095] "Real-time tracking" refers to the process of instantly recording and retaining viewer actions as data the moment they occur.

[0096] In the system implementing this invention, the server captures viewer activity data in real time and preprocesses the data to prepare it for analysis. The server then uses machine learning to extract behavioral pattern features from the captured data and evaluates the viewer's level of interest. Based on this, personalized content suggestions are made in real time.

[0097] The hardware includes smartphones and tablet devices used by users, which have means of communication to send viewing data to the server. The software uses programs such as Python and TensorFlow for data preprocessing and analysis using machine learning models.

[0098] As a concrete example, when a user watches a particular promotional video, actions such as playing, pausing, and rewinding are recorded. The server analyzes this viewing data to identify sections that the user has watched repeatedly and provides the user with additional content related to those sections. This allows users to delve deeper into the topics that interest them.

[0099] An example of a prompt to a generative AI model would be, "Analyze the content of the most viewed sections and generate additional content to recommend to the user." Using this prompt, the AI ​​model can demonstrate its ability to create valuable content based on viewing data.

[0100] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0101] Step 1:

[0102] The device captures user viewing data. It records viewing time, playback, pause, skip, and other events in real time and sends this data to the server. The input is the user's action events, and the output is the viewing event data sent to the server.

[0103] Step 2:

[0104] The server receives and preprocesses the viewing data. It performs data cleaning and converts it into a format that can be processed by machine learning models. The input is viewing event data from the terminal, and the output is formatted viewing data.

[0105] Step 3:

[0106] The server extracts behavioral pattern features from formatted data. It extracts statistical features from viewing data and processes them to quantify interest trends. The input is pre-processed viewing data, and the output is the features of the viewers' behavioral patterns.

[0107] Step 4:

[0108] Based on the features extracted by the server, a generative AI model is used to predict the user's level of interest and appropriate content. Prompt messages are input to the AI ​​model, and prediction results are generated. The input consists of behavioral pattern features and prompt messages, while the output is the level of interest and recommended content.

[0109] Step 5:

[0110] The server provides personalized content to the device based on prediction results. It takes the user's viewing history into consideration and prioritizes suggesting content that might be of interest. The input is the prediction result, and the output is a list of recommended content for the device.

[0111] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0112] This invention is a system that analyzes viewer viewing activity data and emotions in real time to help optimize marketing activities. The system includes functions for data collection, emotion recognition, data analysis, predictive model generation, and result feedback.

[0113] First, the server collects viewing activity data in real time from users watching the product presentation. This includes data such as viewing time, playback / pause actions, and skipped sections. In parallel, the user's video and audio data is fed into the emotion engine for emotion recognition. The emotion engine identifies emotional states such as joy, surprise, and dissatisfaction through facial expression analysis, voice tone analysis, and other methods.

[0114] Next, the server integrates the collected viewing activity data with user sentiment data and extracts features. This allows for the quantification of not only viewers' interests and concerns, but also their intuitive responses based on emotions. This data is then fed into a machine learning model to predict the likelihood of inquiries and applications with high accuracy.

[0115] The server then receives the results estimated by the machine learning model and generates a strategic report for improving marketing activities. This report visualizes changes in viewer sentiment and the resulting distribution of interests, and is provided to the device. This allows marketers to develop audience-optimized strategies from the perspective of video content.

[0116] For example, if a user experiences surprise while watching a video explaining a specific feature of a product, the server will specifically highlight and analyze the data from that moment. The marketing team can then implement a PDCA cycle to stimulate potential purchase intent by reinforcing the elements that evoked surprise and expanding upon them in subsequent videos.

[0117] The following describes the processing flow.

[0118] Step 1:

[0119] The server collects real-time activity data from viewers (users) during the product presentation video, including viewing time, playback, pause, and skipping. Simultaneously, it captures video and audio data in real time via cameras and microphones and supplies this data to the emotion engine.

[0120] Step 2:

[0121] The server uses an emotion engine to recognize the user's emotions from the captured video and audio data. The emotion engine analyzes changes in facial expressions and tone of voice to identify emotions such as joy, surprise, dissatisfaction, and boredom. This allows the user's emotional progression during viewing to be stored as data.

[0122] Step 3:

[0123] The server integrates viewing activity data and emotion recognition data to extract features. These features include high interest in specific video segments, frequency of emotion changes, and viewing time distribution for each type of emotion. This allows for a comprehensive understanding of user behavior patterns and emotional responses.

[0124] Step 4:

[0125] The server uses a machine learning model based on extracted features to predict the likelihood of inquiries or applications. The trained model performs inferences based on historical data and calculates the probability of each user responding.

[0126] Step 5:

[0127] The server analyzes the prediction results and generates a visually organized strategic report. This report includes marketing suggestions, such as the points that viewers were most interested in and the moments when their emotions shifted significantly.

[0128] Step 6:

[0129] The terminal (used by the marketing team) receives reports from the server and uses them to create marketing strategies based on viewer sentiment and behavior data. This helps optimize future video content and design campaigns.

[0130] (Example 2)

[0131] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0132] Conventional audience data analysis systems have difficulty understanding viewers' emotional states, and marketing strategies that take viewers' intuitive reactions into account have not been adequately developed. As a result, content optimization and the accurate capture of viewers' interests have been hindered, limiting the effectiveness of marketing activities.

[0133] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0134] This invention includes a server that captures and preprocesses viewer activity and emotional data in real time, extracts behavioral patterns and emotional state features from the captured data, and utilizes a machine learning model that predicts responses to inquiries and applications based on the extracted features. This enables the optimization of content that reflects the emotional state of viewers and the formulation of highly accurate marketing strategies.

[0135] "Viewer activity data" refers to information about viewers' behavior when viewing content, including viewing time, playback, pausing, skipping, and other events.

[0136] "Emotional data" refers to information that represents the emotional state of the viewer, and includes data that identifies emotional states such as joy, surprise, and dissatisfaction based on facial expressions and tone of voice.

[0137] "Means of preprocessing" refers to a mechanism that processes viewer activity data and emotional data to prepare them for analysis.

[0138] A "method for extracting features" is a mechanism for identifying important patterns and characteristics from the data being analyzed and expressing them numerically or otherwise.

[0139] A "machine learning model" is an algorithm or mathematical model that learns patterns and relationships from data to predict future trends and outcomes.

[0140] A "strategic report" is a strategic document generated based on collected and analyzed data, intended to improve marketing activities.

[0141] "Visualization" refers to depicting and representing complex data and information in an easily understandable way, primarily using graphs and charts.

[0142] This system aims to optimize marketing strategies by collecting and analyzing viewer activity and emotional data in real time. The server collects activity data such as viewing time, playback, pause, and skipping while users are watching product presentation videos. In addition, it uses input devices such as cameras and microphones to capture the user's facial expressions and voice in real time for the emotional analysis engine. Emotional analysis includes facial expression analysis and voice tone analysis, using common analysis software to identify emotional states.

[0143] The server integrates this collected data and extracts features using a programming language like Python and the pandas library. In this process, viewer behavior and emotional patterns are represented as numerical data and formatted into a format that can be input into machine learning frameworks such as TensorFlow and Scikit-learn. Next, the server uses a pre-trained machine learning model based on these features to predict the likelihood of inquiries and applications. The machine learning model analyzes viewer behavior and emotional trends based on previously accumulated data and predicts future behavior with high accuracy.

[0144] Furthermore, the server automatically generates strategic reports utilizing the prediction results of machine learning models. These reports visually represent changes in viewer emotions and viewing data, and are provided to the terminal, making them easily understandable for marketing personnel. This allows them to quickly formulate optimal marketing strategies tailored to the emotional state of the viewers.

[0145] For example, a user might react with surprise when a specific feature of a product is explained to them. In this case, the server highlights the feature data from that moment, revealing key areas for improvement for the marketing team. By emphasizing the elements that evoked this surprise in the next video, the viewer's interest can be further stimulated.

[0146] An example of a prompt is, "How can we identify specific moments that surprised viewers during a product presentation?" This allows for the strategic improvement of video content by leveraging viewer emotions.

[0147] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0148] Step 1:

[0149] The user watches a product presentation video. The server collects user viewing activity data during this process. Data such as viewing time, playback, pause, and skipping are obtained as input. This input data is stored on the server and structured in preparation for analysis.

[0150] Step 2:

[0151] The server collects the user's facial and voice data in real time and feeds it into the emotion recognition engine. Specifically, it uses data obtained from the camera and microphone as input to identify emotional states such as joy, surprise, and dissatisfaction using facial expression analysis and voice tone analysis. As a result of this processing, data on the emotional state is output.

[0152] Step 3:

[0153] The server integrates viewing activity data and sentiment data. Using the Python pandas library, it extracts features from both datasets. Feature extraction yields data quantifying changes in viewers' interest, attention, and sentiment. This data is used as input for machine learning processing.

[0154] Step 4:

[0155] The server inputs feature data into TensorFlow or Scikit-learn and analyzes it using a pre-trained machine learning model. The input is feature data, and the output generates predictions for inquiries and applications. These prediction results are then used to develop marketing strategies.

[0156] Step 5:

[0157] The server generates a strategic report based on the forecast results. The report includes changes in audience sentiment and the resulting trends in viewing activity. The report is sent to terminals and made accessible to marketers. This allows for the development of optimized content strategies for audiences based on visualized data.

[0158] (Application Example 2)

[0159] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0160] Traditional advertising suffers from insufficient detailed analysis of viewers' emotions and behavior, making it difficult to formulate effective marketing strategies. To solve this problem, a method is needed that optimizes advertising using real-time data obtained from viewers' visual and auditory experiences.

[0161] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0162] In this invention, the server includes means for collecting and processing visual and audio data in real time, means for deriving behavioral patterns and emotional states from the collected data, and means for utilizing machine learning models to predict responses based on the derived features and states. This enables the optimization of advertisements based on the emotions and behavior of viewers.

[0163] "Visual data" refers to digital information that captures visual information such as viewers' gaze, facial expressions, and actions in real time.

[0164] "Audio data" refers to digital information that captures auditory information, such as the listener's voice and tone of voice.

[0165] "Means for collecting and processing data in real time" refers to technologies that perform the process of instantly acquiring digital data and converting it into an analyzable format.

[0166] "Behavioral patterns" refer to the regularity of the visual and behavioral behaviors of viewers when they view advertisements.

[0167] "Emotional state" refers to the psychological response of viewers, identifying emotions such as joy, anger, sadness, and happiness.

[0168] A "machine learning model" is an algorithm that learns specific patterns from large amounts of data and uses them to predict future reactions and behaviors.

[0169] "Ad optimization" is the process of adjusting the content and presentation of advertisements based on the interests and concerns of the audience in order to maximize their effectiveness.

[0170] This invention realizes a system that processes visual and audio data in real time to analyze the emotional state and behavioral patterns of viewers. It is implemented according to the following configuration.

[0171] The server collects and processes visual and audio data obtained through smart devices such as Oculus Quest in real time. This includes technologies for user eye tracking, facial recognition, and voice analysis. Specifically, it uses VisageSDK and OpenCV for facial expression analysis and Face API and EigenFace to identify microexpressions. This allows the server to identify the user's emotional state.

[0172] Furthermore, the servers process the collected data in a cloud environment. AWS® Lambda or Azure® ML is used to extract features from the data, and Scikit-learn is used to train machine learning models. These models are used to predict viewer behavior and contribute to advertising optimization.

[0173] The device provides users with feedback to optimize advertising content based on the server's processing results. For example, if a smile is detected while a user is watching an advertisement for sports shoes, the data from that moment is highlighted and analyzed to clearly indicate which elements captured the user's attention.

[0174] For example, if a user shows a strong emotional reaction to a particular product or service while watching an advertisement, that information is used to improve marketing activities. This result is used by a generative AI model to create efficient prompts, presented in text format such as, "Tell us about your reaction to watching this shoe advertisement. What aspects particularly interested you or surprised you?"

[0175] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0176] Step 1:

[0177] Users view advertisements using smart devices. The devices collect visual and audio data in real time. Inputs include the user's gaze, facial expressions, and audio information, while outputs are processable digital data.

[0178] Step 2:

[0179] The server preprocesses the collected data using VisageSDK and OpenCV. In this step, it performs eye tracking and face recognition using the raw input data, converting the user's facial expressions and gaze direction into digital data. The output consists of facial expression data and gaze data.

[0180] Step 3:

[0181] The server utilizes Face API and EigenFace to analyze the user's emotional state from facial expression data. The input is facial expression data, and the output is data indicating the user's emotional state. Here, the server identifies whether the user's emotion is laughter, surprise, or interest.

[0182] Step 4:

[0183] Emotional state data and behavioral pattern data are integrated on a server, and feature extraction is performed. Using a cloud environment, the integrated input data is analyzed, and features indicating specific emotions and behavioral patterns related to advertising are extracted. The output is specific features designed to enhance advertising effectiveness.

[0184] Step 5:

[0185] The server uses Scikit-learn to train a machine learning model and predict future audience reactions. The input is feature data, and the output is the model's result predicting future audience reactions.

[0186] Step 6:

[0187] The device optimizes advertising content based on prediction results generated by the server. It displays ads that respond to the user's emotions and behavior to maximize effectiveness. Here, ad display is adjusted according to the user's specific emotions and actions.

[0188] Step 7:

[0189] The server uses a generation AI model to create prompt messages. The input is user response data, and the output is the prompt message provided to the user. The generated message will be in the form of text such as, "Please tell us about your reaction to watching this shoe advertisement. What aspects particularly interested you or surprised you?"

[0190] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0191] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0192] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0193] [Second Embodiment]

[0194] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0195] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0196] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0197] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0198] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0199] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0200] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0201] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0202] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0203] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0204] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0205] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0206] This invention is a system for analyzing viewer activity data and optimizing marketing activities based on that data. This system is broadly composed of the following processes: data collection, data analysis, predictive model generation, and result feedback.

[0207] First, the server collects diverse data in real time from viewers of the product presentation video. It captures behavioral data such as viewing time, play / pause events, and skipped sections, and converts it into a format that can be immediately analyzed. This process prepares the foundational data necessary to understand the behavioral characteristics of the viewers.

[0208] Next, the server performs analysis based on the collected data. It extracts viewer behavior patterns as features and uses them to run a machine learning model to predict the likelihood of inquiries or applications. This makes it possible to numerically visualize how interested viewers are in the product. This model is designed to make highly accurate predictions by utilizing past training data.

[0209] The server then analyzes the prediction results and generates a strategic report in a format that the marketing team can use. This report includes content that resonated with viewers, responses by segment, and areas for improvement, and is provided to each user's device. This makes it easier for the marketing team to make decisions regarding adjusting video content and executing their next campaign strategies.

[0210] For example, when a user watches a video, sections where they frequently rewind or watch for extended periods are recorded. The server analyzes this data to determine which sections strongly capture the viewer's interest. This information is reflected in reports and used by the marketing team as valuable insights to create new promotional videos that highlight those sections.

[0211] The following describes the processing flow.

[0212] Step 1:

[0213] The server monitors the activity of users watching the product presentation videos in real time and collects data from viewers. Specifically, it acquires and records detailed event data such as the start time of viewing, the timing of playback and stopping, the sections skipped, and the viewing completion rate.

[0214] Step 2:

[0215] The server performs preprocessing on the collected raw data. Preprocessing includes organizing the data, interpolating missing values, and removing noisy data. This process creates a well-organized dataset that is suitable for subsequent analysis.

[0216] Step 3:

[0217] The server extracts viewer behavior patterns as features from the pre-processed data. In this step, multidimensional features such as average viewing time, rewind frequency, and skipping tendencies are calculated. This quantifies viewer interest and engagement.

[0218] Step 4:

[0219] The server uses a machine learning model based on extracted features to predict the likelihood of inquiries or applications. The machine learning model is pre-trained to achieve precise predictions. The model's output represents the probability of a response corresponding to each viewing session.

[0220] Step 5:

[0221] The server receives the prediction results and generates a visual strategic report. This report includes interesting points about viewing patterns and areas for improvement, providing concrete suggestions for marketing activities.

[0222] Step 6:

[0223] The terminal receives strategic reports provided by the server, which are then reviewed by members of the marketing team. This helps them optimize their video content editing and distribution strategies for the future.

[0224] (Example 1)

[0225] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0226] When building effective marketing strategies using viewer behavior data, it is essential to accurately understand viewer interests and efficiently optimize strategies. However, currently, the process from data collection to analysis and strategy proposal is fragmented, making it difficult to respond quickly and make highly accurate predictions.

[0227] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0228] In this invention, the server includes means for collecting and preprocessing viewer behavior information from terminals in real time, means for extracting feature quantities that indicate behavioral trends from the collected behavior information, and means for utilizing a generative AI model that predicts responses to inquiries and suggestions based on the extracted feature quantities. This enables a rapid and highly accurate understanding of viewer interests and the optimization of effective marketing strategies.

[0229] "Device" refers to an electronic device used by viewers to watch video, and includes personal computers, smartphones, tablets, and other similar devices.

[0230] "Viewers" refers to individuals or groups who watch video content and whose viewing behavior is collected as data.

[0231] "Action information" refers to data about the actions viewers perform while watching a video, and specifically includes viewing time, playback, pause, skip, and other events.

[0232] "Collecting data in real time" refers to the process of instantly sending viewer behavior information to the server without delay after an action occurs and acquiring it as data.

[0233] "Preprocessing" refers to the data manipulation required to prepare collected raw data for analysis, and includes imputing missing values, removing outliers, and converting to an appropriate data format.

[0234] In data analysis, "features" refer to data that quantifies patterns or statistical properties used by models for learning and prediction.

[0235] A "generative AI model" refers to an artificial intelligence model that is built to extract regularities and patterns from given data and to make predictions and classifications based on new data.

[0236] "Market research activities" refer to a series of analyses and strategy-building processes conducted to understand market trends for a particular product or service.

[0237] A "strategic report" is a document used to consider the next marketing steps based on the results of market research activities, and refers to a report that includes data analysis results and recommended measures.

[0238] This invention relates to a data analysis system for optimizing marketing activities by utilizing viewer behavior data. This system uses a device on which users view video content and collects their viewing behavior as data. The device consists of personal computers, smartphones, tablets, etc. A server collects viewer behavior information from these devices in real time. This behavior information includes viewing time, playback, pause, skip, and other events, and a JavaScript code snippet runs on the device to send the data to the server.

[0239] Subsequently, the server preprocesses the collected data, performing tasks such as imputing missing values ​​and removing outliers, and converting the data into a unified format using libraries such as Python's Pandas. From the preprocessed data, the server extracts features of viewer behavior patterns and prepares them as input data for a machine learning model. The generative AI model utilizes a pre-trained neural network using libraries such as Scikit-learn and TensorFlow.

[0240] The server uses a generative AI model to predict viewers' levels of interest, inquiries, and responses to suggestions. This makes it possible to quantify and visualize the degree to which viewers are interested in a product or service. Based on these predictions, the server generates a strategic report for market research activities and provides it to each remote terminal.

[0241] For example, if a user repeatedly plays a specific section while watching a product explanation video, the server will determine that the user has a high level of interest in that section and reflect this information in the report. The marketing team can then use this report to create new promotional content that highlights that section.

[0242] An example of a prompt would be: "Analyze the predictive report generated from viewing data and propose a new promotional strategy that highlights the sections that particularly captured audience interest."

[0243] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0244] Step 1:

[0245] The device collects real-time information about the user's actions while watching videos. This information includes viewing time, playback, pause, skipping, etc. A JavaScript code snippet is used to build a system that periodically sends this data to the server. The input to this process is the user's viewing behavior, and the output is the behavioral data sent to the server.

[0246] Step 2:

[0247] The server preprocesses the operational information received from the terminal. Specifically, it uses the Python Pandas library to impute missing values ​​and filter out outliers. Afterward, it converts the data into a unified format, preparing it for analysis. The input for this step is the raw data received from the terminal, and the output is the preprocessed, purified data.

[0248] Step 3:

[0249] The server extracts features from pre-processed data. It quantifies viewer behavior patterns, such as the distribution of viewing time and the proportion of frequently played sections. Statistical methods and data analysis techniques are used for this feature extraction. The input for this step is pre-processed data, and the output is a feature vector.

[0250] Step 4:

[0251] The server processes feature vectors using a generative AI model to predict viewer interest and likelihood of response. It then applies a pre-trained model using Scikit-learn or TensorFlow to calculate a prediction score. The input for this step is a feature vector, and the output is the prediction score.

[0252] Step 5:

[0253] The server generates a strategic report on market research activities based on the prediction score. Visualization tools are used to format the report in a visually easy-to-understand manner. The report includes content sections that viewers showed particular interest in, as well as recommended improvement actions. The input for this step is the prediction score, and the output is the strategic report.

[0254] Step 6:

[0255] The user (marketing team) uses the generated strategic report to optimize their next marketing strategy. Based on the feedback in the report, they implement specific measures, such as creating new promotional content. The input for this step is the strategic report, and the output is the improved marketing activities.

[0256] (Application Example 1)

[0257] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0258] Traditional methods of utilizing viewer data have made it difficult to efficiently analyze large amounts of data and formulate accurate marketing strategies based on that analysis. Furthermore, providing content that reflects individual viewer interests and behavioral patterns in real time has been challenging, resulting in difficulties in improving customer satisfaction and business performance.

[0259] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0260] In this invention, the server includes means for capturing and pre-processing viewer activity data in real time, means for extracting behavioral pattern features from the captured data, and means for tracking viewer responses to each viewing section in real time and providing personalized content suggestions. This enables direct utilization of viewer reactions to provide dynamic content tailored to customer interests and optimize marketing strategies.

[0261] "Activity data" refers to data about viewers' actions when viewing content, including viewing time and events such as playback, pausing, and skipping.

[0262] "Preprocessing" refers to the process of preparing data into an analyzable format, including tasks such as data cleaning and format conversion.

[0263] "Behavioral pattern features" are analytical indicators extracted from viewer behavior data, and are elements used to quantify viewing trends and interests.

[0264] "Personalized content recommendations" is an approach that presents specific content tailored to an individual's interests based on their behavioral data.

[0265] "Real-time tracking" refers to the process of instantly recording and retaining viewer actions as data the moment they occur.

[0266] In the system implementing this invention, the server captures viewer activity data in real time and preprocesses the data to prepare it for analysis. The server then uses machine learning to extract behavioral pattern features from the captured data and evaluates the viewer's level of interest. Based on this, personalized content suggestions are made in real time.

[0267] The hardware includes smartphones and tablet devices used by users, which have means of communication to send viewing data to the server. The software uses programs such as Python and TensorFlow for data preprocessing and analysis using machine learning models.

[0268] As a concrete example, when a user watches a particular promotional video, actions such as playing, pausing, and rewinding are recorded. The server analyzes this viewing data to identify sections that the user has watched repeatedly and provides the user with additional content related to those sections. This allows users to delve deeper into the topics that interest them.

[0269] An example of a prompt to a generative AI model would be, "Analyze the content of the most viewed sections and generate additional content to recommend to the user." Using this prompt, the AI ​​model can demonstrate its ability to create valuable content based on viewing data.

[0270] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0271] Step 1:

[0272] The device captures user viewing data. It records viewing time, playback, pause, skip, and other events in real time and sends this data to the server. The input is the user's action events, and the output is the viewing event data sent to the server.

[0273] Step 2:

[0274] The server receives and preprocesses the viewing data. It performs data cleaning and converts it into a format that can be processed by machine learning models. The input is viewing event data from the terminal, and the output is formatted viewing data.

[0275] Step 3:

[0276] The server extracts behavioral pattern features from formatted data. It extracts statistical features from viewing data and processes them to quantify interest trends. The input is pre-processed viewing data, and the output is the features of the viewers' behavioral patterns.

[0277] Step 4:

[0278] Based on the features extracted by the server, a generative AI model is used to predict the user's level of interest and appropriate content. Prompt messages are input to the AI ​​model, and prediction results are generated. The input consists of behavioral pattern features and prompt messages, while the output is the level of interest and recommended content.

[0279] Step 5:

[0280] The server provides the terminal with personalized content based on the prediction results. Considering the user's viewing history, it prioritizes and proposes content that is appealing. The input is the prediction result, and the output is a list of recommended content for the terminal.

[0281] Furthermore, an emotion engine for estimating the user's emotions may be combined. That is, the specific processing unit 290 may estimate the user's emotions using the emotion recognition model 59 and perform specific processing using the user's emotions.

[0282] The present invention is a system that analyzes the viewing activity data and emotions of viewers in real time and is useful for optimizing marketing activities. The system includes functions for data collection, emotion recognition, data analysis, generation of a prediction model, and result feedback.

[0283] First, the server collects viewing activity data in real time from users who are watching a product presentation. This includes data such as viewing time, play / stop operations, and skipped parts. In parallel, the user's video and audio data are fed into the emotion engine for emotion recognition. The emotion engine identifies emotional states such as joy, surprise, and dissatisfaction through facial expression analysis, voice tone analysis, etc.

[0284] Next, the server integrates the collected viewing activity data and the user's emotion data and extracts features. This enables quantification not only of the viewer's interests and concerns but also of intuitive reactions based on emotions. These data are input into a machine learning model for accurately predicting the likelihood of inquiries and applications.

[0285] After that, the server receives the results estimated by the machine learning model and generates a strategic report for improving marketing activities. This report visualizes the changes in viewers' emotions and the accompanying interest distribution and is provided to the terminal. As a result, marketing personnel can formulate strategies optimized for viewers from the perspective of video content.

[0286] As a specific example, when a surprise emotion is recognized when a user watches an explanation of a specific function of a product, the server analyzes the data at that moment with particular emphasis. The marketing team can implement PDCA to stimulate potential purchasing desire by strengthening the elements that induced surprise and further developing them in the next video.

[0287] The following explains the processing flow.

[0288] Step 1:

[0289] During the viewing of the product presentation video, the server collects in real time the activity data of the viewer (user), such as viewing time, play, stop, and skip. In parallel, video and audio data are also captured in real time through the camera and microphone and supplied to the emotion engine.

[0290] Step 2:

[0291] The server uses the emotion engine to recognize the user's emotions from the captured video and audio data. The emotion engine analyzes changes in facial expressions and tones of voice to identify emotions such as joy, surprise, dissatisfaction, and boredom. As a result, the transition of the user's emotions during viewing is saved as data.

[0292] Step 3:

[0293] The server integrates viewing activity data and emotion recognition data to extract features. These features include high interest in specific video segments, frequency of emotion changes, and viewing time distribution for each type of emotion. This allows for a comprehensive understanding of user behavior patterns and emotional responses.

[0294] Step 4:

[0295] The server uses a machine learning model based on extracted features to predict the likelihood of inquiries or applications. The trained model performs inferences based on historical data and calculates the probability of each user responding.

[0296] Step 5:

[0297] The server analyzes the prediction results and generates a visually organized strategic report. This report includes marketing suggestions, such as the points that viewers were most interested in and the moments when their emotions shifted significantly.

[0298] Step 6:

[0299] The terminal (used by the marketing team) receives reports from the server and uses them to create marketing strategies based on viewer sentiment and behavior data. This helps optimize future video content and design campaigns.

[0300] (Example 2)

[0301] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0302] Conventional audience data analysis systems have difficulty understanding viewers' emotional states, and marketing strategies that take viewers' intuitive reactions into account have not been adequately developed. As a result, content optimization and the accurate capture of viewers' interests have been hindered, limiting the effectiveness of marketing activities.

[0303] The specific processing by the specific processing unit 290 of the data processing apparatus 12 in Example 2 is realized by the following means.

[0304] In this invention, the server includes means for capturing viewer activity data and emotion data in real time and performing preprocessing, means for extracting feature quantities of behavior patterns and emotion states from the captured data, and means for utilizing a machine learning model for predicting responses to inquiries and applications based on the extracted feature quantities. Thereby, it becomes possible to optimize content reflecting the viewer's emotion state and formulate a highly accurate marketing strategy.

[0305] "Viewer activity data" is information regarding the behavior of a viewer when viewing content, and is data including events such as viewing time, play, stop, skip, etc.

[0306] "Emotion data" is information representing the emotion state of a viewer, and is data including emotion states such as joy, surprise, dissatisfaction, etc. identified based on expressions and tones of voice.

[0307] "Means for performing preprocessing" is a mechanism for performing processing to arrange viewer activity data and emotion data in an analyzable format.

[0308] "Means for extracting feature quantities" is a mechanism for finding important patterns and characteristics from the data to be analyzed and expressing them in numerical values or the like.

[0309] "Machine learning model" is an algorithm or mathematical model for learning patterns and correlations from data and predicting future trends and results.

[0310] "Strategy report" is a strategic document generated based on the collected and analyzed data for the purpose of improving marketing activities.

[0311] "Visualization" refers to depicting and representing complex data and information in an easily understandable way, primarily using graphs and charts.

[0312] This system aims to optimize marketing strategies by collecting and analyzing viewer activity and emotional data in real time. The server collects activity data such as viewing time, playback, pause, and skipping while users are watching product presentation videos. In addition, it uses input devices such as cameras and microphones to capture the user's facial expressions and voice in real time for the emotional analysis engine. Emotional analysis includes facial expression analysis and voice tone analysis, using common analysis software to identify emotional states.

[0313] The server integrates this collected data and extracts features using a programming language like Python and the pandas library. In this process, viewer behavior and emotional patterns are represented as numerical data and formatted into a format that can be input into machine learning frameworks such as TensorFlow and Scikit-learn. Next, the server uses a pre-trained machine learning model based on these features to predict the likelihood of inquiries and applications. The machine learning model analyzes viewer behavior and emotional trends based on previously accumulated data and predicts future behavior with high accuracy.

[0314] Furthermore, the server automatically generates strategic reports utilizing the prediction results of machine learning models. These reports visually represent changes in viewer emotions and viewing data, and are provided to the terminal, making them easily understandable for marketing personnel. This allows them to quickly formulate optimal marketing strategies tailored to the emotional state of the viewers.

[0315] For example, a user might react with surprise when a specific feature of a product is explained to them. In this case, the server highlights the feature data from that moment, revealing key areas for improvement for the marketing team. By emphasizing the elements that evoked this surprise in the next video, the viewer's interest can be further stimulated.

[0316] An example of a prompt is, "How can we identify specific moments that surprised viewers during a product presentation?" This allows for the strategic improvement of video content by leveraging viewer emotions.

[0317] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0318] Step 1:

[0319] The user watches a product presentation video. The server collects user viewing activity data during this process. Data such as viewing time, playback, pause, and skipping are obtained as input. This input data is stored on the server and structured in preparation for analysis.

[0320] Step 2:

[0321] The server collects the user's facial and voice data in real time and feeds it into the emotion recognition engine. Specifically, it uses data obtained from the camera and microphone as input to identify emotional states such as joy, surprise, and dissatisfaction using facial expression analysis and voice tone analysis. As a result of this processing, data on the emotional state is output.

[0322] Step 3:

[0323] The server integrates viewing activity data and sentiment data. Using the Python pandas library, it extracts features from both datasets. Feature extraction yields data quantifying changes in viewers' interest, attention, and sentiment. This data is used as input for machine learning processing.

[0324] Step 4:

[0325] The server inputs feature data into TensorFlow or Scikit-learn and analyzes it using a pre-trained machine learning model. The input is feature data, and the output generates predictions for inquiries and applications. These prediction results are then used to develop marketing strategies.

[0326] Step 5:

[0327] The server generates a strategic report based on the forecast results. The report includes changes in audience sentiment and the resulting trends in viewing activity. The report is sent to terminals and made accessible to marketers. This allows for the development of optimized content strategies for audiences based on visualized data.

[0328] (Application Example 2)

[0329] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0330] Traditional advertising suffers from insufficient detailed analysis of viewers' emotions and behavior, making it difficult to formulate effective marketing strategies. To solve this problem, a method is needed that optimizes advertising using real-time data obtained from viewers' visual and auditory experiences.

[0331] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0332] In this invention, the server includes means for collecting and processing visual and audio data in real time, means for deriving behavioral patterns and emotional states from the collected data, and means for utilizing machine learning models to predict responses based on the derived features and states. This enables the optimization of advertisements based on the emotions and behavior of viewers.

[0333] "Visual data" refers to digital information that captures visual information such as viewers' gaze, facial expressions, and actions in real time.

[0334] "Audio data" refers to digital information that captures auditory information, such as the listener's voice and tone of voice.

[0335] "Means for collecting and processing data in real time" refers to technologies that perform the process of instantly acquiring digital data and converting it into an analyzable format.

[0336] "Behavioral patterns" refer to the regularity of the visual and behavioral behaviors of viewers when they view advertisements.

[0337] "Emotional state" refers to the psychological response of viewers, identifying emotions such as joy, anger, sadness, and happiness.

[0338] A "machine learning model" is an algorithm that learns specific patterns from large amounts of data and uses them to predict future reactions and behaviors.

[0339] "Ad optimization" is the process of adjusting the content and presentation of advertisements based on the interests and concerns of the audience in order to maximize their effectiveness.

[0340] This invention realizes a system that processes visual and audio data in real time to analyze the emotional state and behavioral patterns of viewers. It is implemented according to the following configuration.

[0341] The server collects and processes visual and audio data obtained through smart devices such as Oculus Quest in real time. This includes technologies for user eye tracking, facial recognition, and voice analysis. Specifically, it uses VisageSDK and OpenCV for facial expression analysis and Face API and EigenFace to identify microexpressions. This allows the server to identify the user's emotional state.

[0342] Furthermore, the servers process the collected data in a cloud environment. AWS Lambda or Azure ML is used to extract features from the data, and Scikit-learn is used to train machine learning models. These models are used to predict viewer behavior and contribute to ad optimization.

[0343] The device provides users with feedback to optimize advertising content based on the server's processing results. For example, if a smile is detected while a user is watching an advertisement for sports shoes, the data from that moment is highlighted and analyzed to clearly indicate which elements captured the user's attention.

[0344] For example, if a user shows a strong emotional reaction to a particular product or service while watching an advertisement, that information is used to improve marketing activities. This result is used by a generative AI model to create efficient prompts, presented in text format such as, "Tell us about your reaction to watching this shoe advertisement. What aspects particularly interested you or surprised you?"

[0345] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0346] Step 1:

[0347] Users view advertisements using smart devices. The devices collect visual and audio data in real time. Inputs include the user's gaze, facial expressions, and audio information, while outputs are processable digital data.

[0348] Step 2:

[0349] The server preprocesses the collected data using VisageSDK and OpenCV. In this step, it performs eye tracking and face recognition using the raw input data, converting the user's facial expressions and gaze direction into digital data. The output consists of facial expression data and gaze data.

[0350] Step 3:

[0351] The server utilizes Face API and EigenFace to analyze the user's emotional state from facial expression data. The input is facial expression data, and the output is data indicating the user's emotional state. Here, the server identifies whether the user's emotion is laughter, surprise, or interest.

[0352] Step 4:

[0353] Emotional state data and behavioral pattern data are integrated on a server, and feature extraction is performed. Using a cloud environment, the integrated input data is analyzed, and features indicating specific emotions and behavioral patterns related to advertising are extracted. The output is specific features designed to enhance advertising effectiveness.

[0354] Step 5:

[0355] The server uses Scikit-learn to train a machine learning model and predict future audience reactions. The input is feature data, and the output is the model's result predicting future audience reactions.

[0356] Step 6:

[0357] The device optimizes advertising content based on prediction results generated by the server. It displays ads that respond to the user's emotions and behavior to maximize effectiveness. Here, ad display is adjusted according to the user's specific emotions and actions.

[0358] Step 7:

[0359] The server uses a generation AI model to create prompt messages. The input is user response data, and the output is the prompt message provided to the user. The generated message will be in the form of text such as, "Please tell us about your reaction to watching this shoe advertisement. What aspects particularly interested you or surprised you?"

[0360] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0361] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0362] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0363] [Third Embodiment]

[0364] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0365] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0366] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0367] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0368] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0369] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0370] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0371] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0372] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0373] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0374] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0375] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0376] This invention is a system for analyzing viewer activity data and optimizing marketing activities based on that data. This system is broadly composed of the following processes: data collection, data analysis, predictive model generation, and result feedback.

[0377] First, the server collects diverse data in real time from viewers of the product presentation video. It captures behavioral data such as viewing time, play / pause events, and skipped sections, and converts it into a format that can be immediately analyzed. This process prepares the foundational data necessary to understand the behavioral characteristics of the viewers.

[0378] Next, the server performs analysis based on the collected data. It extracts viewer behavior patterns as features and uses them to run a machine learning model to predict the likelihood of inquiries or applications. This makes it possible to numerically visualize how interested viewers are in the product. This model is designed to make highly accurate predictions by utilizing past training data.

[0379] The server then analyzes the prediction results and generates a strategic report in a format that the marketing team can use. This report includes content that resonated with viewers, responses by segment, and areas for improvement, and is provided to each user's device. This makes it easier for the marketing team to make decisions regarding adjusting video content and executing their next campaign strategies.

[0380] For example, when a user watches a video, sections where they frequently rewind or watch for extended periods are recorded. The server analyzes this data to determine which sections strongly capture the viewer's interest. This information is reflected in reports and used by the marketing team as valuable insights to create new promotional videos that highlight those sections.

[0381] The following describes the processing flow.

[0382] Step 1:

[0383] The server monitors the activity of users watching the product presentation videos in real time and collects data from viewers. Specifically, it acquires and records detailed event data such as the start time of viewing, the timing of playback and stopping, the sections skipped, and the viewing completion rate.

[0384] Step 2:

[0385] The server performs preprocessing on the collected raw data. Preprocessing includes organizing the data, interpolating missing values, and removing noisy data. This process creates a well-organized dataset that is suitable for subsequent analysis.

[0386] Step 3:

[0387] The server extracts viewer behavior patterns as features from the pre-processed data. In this step, multidimensional features such as average viewing time, rewind frequency, and skipping tendencies are calculated. This quantifies viewer interest and engagement.

[0388] Step 4:

[0389] The server uses a machine learning model based on extracted features to predict the likelihood of inquiries or applications. The machine learning model is pre-trained to achieve precise predictions. The model's output represents the probability of a response corresponding to each viewing session.

[0390] Step 5:

[0391] The server receives the prediction results and generates a visual strategic report. This report includes interesting points about viewing patterns and areas for improvement, providing concrete suggestions for marketing activities.

[0392] Step 6:

[0393] The terminal receives strategic reports provided by the server, which are then reviewed by members of the marketing team. This helps them optimize their video content editing and distribution strategies for the future.

[0394] (Example 1)

[0395] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0396] When building effective marketing strategies using viewer behavior data, it is essential to accurately understand viewer interests and efficiently optimize strategies. However, currently, the process from data collection to analysis and strategy proposal is fragmented, making it difficult to respond quickly and make highly accurate predictions.

[0397] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0398] In this invention, the server includes means for collecting and preprocessing viewer behavior information from terminals in real time, means for extracting feature quantities that indicate behavioral trends from the collected behavior information, and means for utilizing a generative AI model that predicts responses to inquiries and suggestions based on the extracted feature quantities. This enables a rapid and highly accurate understanding of viewer interests and the optimization of effective marketing strategies.

[0399] "Device" refers to an electronic device used by viewers to watch video, and includes personal computers, smartphones, tablets, and other similar devices.

[0400] "Viewers" refers to individuals or groups who watch video content and whose viewing behavior is collected as data.

[0401] "Action information" refers to data about the actions viewers perform while watching a video, and specifically includes viewing time, playback, pause, skip, and other events.

[0402] "Collecting data in real time" refers to the process of instantly sending viewer behavior information to the server without delay after an action occurs and acquiring it as data.

[0403] "Preprocessing" refers to the data manipulation required to prepare collected raw data for analysis, and includes imputing missing values, removing outliers, and converting to an appropriate data format.

[0404] In data analysis, "features" refer to data that quantifies patterns or statistical properties used by models for learning and prediction.

[0405] A "generative AI model" refers to an artificial intelligence model that is built to extract regularities and patterns from given data and to make predictions and classifications based on new data.

[0406] "Market research activities" refer to a series of analyses and strategy-building processes conducted to understand market trends for a particular product or service.

[0407] A "strategic report" is a document used to consider the next marketing steps based on the results of market research activities, and refers to a report that includes data analysis results and recommended measures.

[0408] This invention relates to a data analysis system for optimizing marketing activities by utilizing viewer behavior data. This system uses a device on which users view video content and collects their viewing behavior as data. The device consists of personal computers, smartphones, tablets, etc. A server collects viewer behavior information from these devices in real time. This behavior information includes viewing time, playback, pause, skip, and other events, and a JavaScript code snippet runs on the device to send the data to the server.

[0409] Subsequently, the server preprocesses the collected data, performing tasks such as imputing missing values ​​and removing outliers, and converting the data into a unified format using libraries such as Python's Pandas. From the preprocessed data, the server extracts features of viewer behavior patterns and prepares them as input data for a machine learning model. The generative AI model utilizes a pre-trained neural network using libraries such as Scikit-learn and TensorFlow.

[0410] The server uses a generative AI model to predict viewers' levels of interest, inquiries, and responses to suggestions. This makes it possible to quantify and visualize the degree to which viewers are interested in a product or service. Based on these predictions, the server generates a strategic report for market research activities and provides it to each remote terminal.

[0411] For example, if a user repeatedly plays a specific section while watching a product explanation video, the server will determine that the user has a high level of interest in that section and reflect this information in the report. The marketing team can then use this report to create new promotional content that highlights that section.

[0412] An example of a prompt would be: "Analyze the predictive report generated from viewing data and propose a new promotional strategy that highlights the sections that particularly captured audience interest."

[0413] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0414] Step 1:

[0415] The device collects real-time information about the user's actions while watching videos. This information includes viewing time, playback, pause, skipping, etc. A JavaScript code snippet is used to build a system that periodically sends this data to the server. The input to this process is the user's viewing behavior, and the output is the behavioral data sent to the server.

[0416] Step 2:

[0417] The server preprocesses the operational information received from the terminal. Specifically, it uses the Python Pandas library to impute missing values ​​and filter out outliers. Afterward, it converts the data into a unified format, preparing it for analysis. The input for this step is the raw data received from the terminal, and the output is the preprocessed, purified data.

[0418] Step 3:

[0419] The server extracts features from pre-processed data. It quantifies viewer behavior patterns, such as the distribution of viewing time and the proportion of frequently played sections. Statistical methods and data analysis techniques are used for this feature extraction. The input for this step is pre-processed data, and the output is a feature vector.

[0420] Step 4:

[0421] The server processes feature vectors using a generative AI model to predict viewer interest and likelihood of response. It then applies a pre-trained model using Scikit-learn or TensorFlow to calculate a prediction score. The input for this step is a feature vector, and the output is the prediction score.

[0422] Step 5:

[0423] The server generates a strategic report on market research activities based on the prediction score. Visualization tools are used to format the report in a visually easy-to-understand manner. The report includes content sections that viewers showed particular interest in, as well as recommended improvement actions. The input for this step is the prediction score, and the output is the strategic report.

[0424] Step 6:

[0425] The user (marketing team) uses the generated strategic report to optimize their next marketing strategy. Based on the feedback in the report, they implement specific measures, such as creating new promotional content. The input for this step is the strategic report, and the output is the improved marketing activities.

[0426] (Application Example 1)

[0427] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0428] Traditional methods of utilizing viewer data have made it difficult to efficiently analyze large amounts of data and formulate accurate marketing strategies based on that analysis. Furthermore, providing content that reflects individual viewer interests and behavioral patterns in real time has been challenging, resulting in difficulties in improving customer satisfaction and business performance.

[0429] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0430] In this invention, the server includes means for capturing and pre-processing viewer activity data in real time, means for extracting behavioral pattern features from the captured data, and means for tracking viewer responses to each viewing section in real time and providing personalized content suggestions. This enables direct utilization of viewer reactions to provide dynamic content tailored to customer interests and optimize marketing strategies.

[0431] "Activity data" refers to data about viewers' actions when viewing content, including viewing time and events such as playback, pausing, and skipping.

[0432] "Preprocessing" refers to the process of preparing data into an analyzable format, including tasks such as data cleaning and format conversion.

[0433] "Behavioral pattern features" are analytical indicators extracted from viewer behavior data, and are elements used to quantify viewing trends and interests.

[0434] "Personalized content recommendations" is an approach that presents specific content tailored to an individual's interests based on their behavioral data.

[0435] "Real-time tracking" refers to the process of instantly recording and retaining viewer actions as data the moment they occur.

[0436] In the system implementing this invention, the server captures viewer activity data in real time and preprocesses the data to prepare it for analysis. The server then uses machine learning to extract behavioral pattern features from the captured data and evaluates the viewer's level of interest. Based on this, personalized content suggestions are made in real time.

[0437] The hardware includes smartphones and tablet devices used by users, which have means of communication to send viewing data to the server. The software uses programs such as Python and TensorFlow for data preprocessing and analysis using machine learning models.

[0438] As a concrete example, when a user watches a particular promotional video, actions such as playing, pausing, and rewinding are recorded. The server analyzes this viewing data to identify sections that the user has watched repeatedly and provides the user with additional content related to those sections. This allows users to delve deeper into the topics that interest them.

[0439] An example of a prompt to a generative AI model would be, "Analyze the content of the most viewed sections and generate additional content to recommend to the user." Using this prompt, the AI ​​model can demonstrate its ability to create valuable content based on viewing data.

[0440] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0441] Step 1:

[0442] The device captures user viewing data. It records viewing time, playback, pause, skip, and other events in real time and sends this data to the server. The input is the user's action events, and the output is the viewing event data sent to the server.

[0443] Step 2:

[0444] The server receives and preprocesses the viewing data. It performs data cleaning and converts it into a format that can be processed by machine learning models. The input is viewing event data from the terminal, and the output is formatted viewing data.

[0445] Step 3:

[0446] The server extracts behavioral pattern features from formatted data. It extracts statistical features from viewing data and processes them to quantify interest trends. The input is pre-processed viewing data, and the output is the features of the viewers' behavioral patterns.

[0447] Step 4:

[0448] Based on the features extracted by the server, a generative AI model is used to predict the user's level of interest and appropriate content. Prompt messages are input to the AI ​​model, and prediction results are generated. The input consists of behavioral pattern features and prompt messages, while the output is the level of interest and recommended content.

[0449] Step 5:

[0450] The server provides personalized content to the device based on prediction results. It takes the user's viewing history into consideration and prioritizes suggesting content that might be of interest. The input is the prediction result, and the output is a list of recommended content for the device.

[0451] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0452] This invention is a system that analyzes viewer viewing activity data and emotions in real time to help optimize marketing activities. The system includes functions for data collection, emotion recognition, data analysis, predictive model generation, and result feedback.

[0453] First, the server collects viewing activity data in real time from users watching the product presentation. This includes data such as viewing time, playback / pause actions, and skipped sections. In parallel, the user's video and audio data is fed into the emotion engine for emotion recognition. The emotion engine identifies emotional states such as joy, surprise, and dissatisfaction through facial expression analysis, voice tone analysis, and other methods.

[0454] Next, the server integrates the collected viewing activity data with user sentiment data and extracts features. This allows for the quantification of not only viewers' interests and concerns, but also their intuitive responses based on emotions. This data is then fed into a machine learning model to predict the likelihood of inquiries and applications with high accuracy.

[0455] The server then receives the results estimated by the machine learning model and generates a strategic report for improving marketing activities. This report visualizes changes in viewer sentiment and the resulting distribution of interests, and is provided to the device. This allows marketers to develop audience-optimized strategies from the perspective of video content.

[0456] For example, if a user experiences surprise while watching a video explaining a specific feature of a product, the server will specifically highlight and analyze the data from that moment. The marketing team can then implement a PDCA cycle to stimulate potential purchase intent by reinforcing the elements that evoked surprise and expanding upon them in subsequent videos.

[0457] The following describes the processing flow.

[0458] Step 1:

[0459] The server collects real-time activity data from viewers (users) during the product presentation video, including viewing time, playback, pause, and skipping. Simultaneously, it captures video and audio data in real time via cameras and microphones and supplies this data to the emotion engine.

[0460] Step 2:

[0461] The server uses an emotion engine to recognize the user's emotions from the captured video and audio data. The emotion engine analyzes changes in facial expressions and tone of voice to identify emotions such as joy, surprise, dissatisfaction, and boredom. This allows the user's emotional progression during viewing to be stored as data.

[0462] Step 3:

[0463] The server integrates viewing activity data and emotion recognition data to extract features. These features include high interest in specific video segments, frequency of emotion changes, and viewing time distribution for each type of emotion. This allows for a comprehensive understanding of user behavior patterns and emotional responses.

[0464] Step 4:

[0465] The server uses a machine learning model based on extracted features to predict the likelihood of inquiries or applications. The trained model performs inferences based on historical data and calculates the probability of each user responding.

[0466] Step 5:

[0467] The server analyzes the prediction results and generates a visually organized strategic report. This report includes marketing suggestions, such as the points that viewers were most interested in and the moments when their emotions shifted significantly.

[0468] Step 6:

[0469] The terminal (used by the marketing team) receives reports from the server and uses them to create marketing strategies based on viewer sentiment and behavior data. This helps optimize future video content and design campaigns.

[0470] (Example 2)

[0471] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0472] Conventional audience data analysis systems have difficulty understanding viewers' emotional states, and marketing strategies that take viewers' intuitive reactions into account have not been adequately developed. As a result, content optimization and the accurate capture of viewers' interests have been hindered, limiting the effectiveness of marketing activities.

[0473] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0474] This invention includes a server that captures and preprocesses viewer activity and emotional data in real time, extracts behavioral patterns and emotional state features from the captured data, and utilizes a machine learning model that predicts responses to inquiries and applications based on the extracted features. This enables the optimization of content that reflects the emotional state of viewers and the formulation of highly accurate marketing strategies.

[0475] "Viewer activity data" refers to information about viewers' behavior when viewing content, including viewing time, playback, pausing, skipping, and other events.

[0476] "Emotional data" refers to information that represents the emotional state of the viewer, and includes data that identifies emotional states such as joy, surprise, and dissatisfaction based on facial expressions and tone of voice.

[0477] "Means of preprocessing" refers to a mechanism that processes viewer activity data and emotional data to prepare them for analysis.

[0478] A "method for extracting features" is a mechanism for identifying important patterns and characteristics from the data being analyzed and expressing them numerically or otherwise.

[0479] A "machine learning model" is an algorithm or mathematical model that learns patterns and relationships from data to predict future trends and outcomes.

[0480] A "strategic report" is a strategic document generated based on collected and analyzed data, intended to improve marketing activities.

[0481] "Visualization" refers to depicting and representing complex data and information in an easily understandable way, primarily using graphs and charts.

[0482] This system aims to optimize marketing strategies by collecting and analyzing viewer activity and emotional data in real time. The server collects activity data such as viewing time, playback, pause, and skipping while users are watching product presentation videos. In addition, it uses input devices such as cameras and microphones to capture the user's facial expressions and voice in real time for the emotional analysis engine. Emotional analysis includes facial expression analysis and voice tone analysis, using common analysis software to identify emotional states.

[0483] The server integrates this collected data and extracts features using a programming language like Python and the pandas library. In this process, viewer behavior and emotional patterns are represented as numerical data and formatted into a format that can be input into machine learning frameworks such as TensorFlow and Scikit-learn. Next, the server uses a pre-trained machine learning model based on these features to predict the likelihood of inquiries and applications. The machine learning model analyzes viewer behavior and emotional trends based on previously accumulated data and predicts future behavior with high accuracy.

[0484] Furthermore, the server automatically generates strategic reports utilizing the prediction results of machine learning models. These reports visually represent changes in viewer emotions and viewing data, and are provided to the terminal, making them easily understandable for marketing personnel. This allows them to quickly formulate optimal marketing strategies tailored to the emotional state of the viewers.

[0485] For example, a user might react with surprise when a specific feature of a product is explained to them. In this case, the server highlights the feature data from that moment, revealing key areas for improvement for the marketing team. By emphasizing the elements that evoked this surprise in the next video, the viewer's interest can be further stimulated.

[0486] An example of a prompt is, "How can we identify specific moments that surprised viewers during a product presentation?" This allows for the strategic improvement of video content by leveraging viewer emotions.

[0487] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0488] Step 1:

[0489] The user watches a product presentation video. The server collects user viewing activity data during this process. Data such as viewing time, playback, pause, and skipping are obtained as input. This input data is stored on the server and structured in preparation for analysis.

[0490] Step 2:

[0491] The server collects the user's facial and voice data in real time and feeds it into the emotion recognition engine. Specifically, it uses data obtained from the camera and microphone as input to identify emotional states such as joy, surprise, and dissatisfaction using facial expression analysis and voice tone analysis. As a result of this processing, data on the emotional state is output.

[0492] Step 3:

[0493] The server integrates viewing activity data and sentiment data. Using the Python pandas library, it extracts features from both datasets. Feature extraction yields data quantifying changes in viewers' interest, attention, and sentiment. This data is used as input for machine learning processing.

[0494] Step 4:

[0495] The server inputs feature data into TensorFlow or Scikit-learn and analyzes it using a pre-trained machine learning model. The input is feature data, and the output generates predictions for inquiries and applications. These prediction results are then used to develop marketing strategies.

[0496] Step 5:

[0497] The server generates a strategic report based on the forecast results. The report includes changes in audience sentiment and the resulting trends in viewing activity. The report is sent to terminals and made accessible to marketers. This allows for the development of optimized content strategies for audiences based on visualized data.

[0498] (Application Example 2)

[0499] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0500] Traditional advertising suffers from insufficient detailed analysis of viewers' emotions and behavior, making it difficult to formulate effective marketing strategies. To solve this problem, a method is needed that optimizes advertising using real-time data obtained from viewers' visual and auditory experiences.

[0501] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0502] In this invention, the server includes means for collecting and processing visual and audio data in real time, means for deriving behavioral patterns and emotional states from the collected data, and means for utilizing machine learning models to predict responses based on the derived features and states. This enables the optimization of advertisements based on the emotions and behavior of viewers.

[0503] "Visual data" refers to digital information that captures visual information such as viewers' gaze, facial expressions, and actions in real time.

[0504] "Audio data" refers to digital information that captures auditory information, such as the listener's voice and tone of voice.

[0505] "Means for collecting and processing data in real time" refers to technologies that perform the process of instantly acquiring digital data and converting it into an analyzable format.

[0506] "Behavioral patterns" refer to the regularity of the visual and behavioral behaviors of viewers when they view advertisements.

[0507] "Emotional state" refers to the psychological response of viewers, identifying emotions such as joy, anger, sadness, and happiness.

[0508] A "machine learning model" is an algorithm that learns specific patterns from large amounts of data and uses them to predict future reactions and behaviors.

[0509] "Ad optimization" is the process of adjusting the content and presentation of advertisements based on the interests and concerns of the audience in order to maximize their effectiveness.

[0510] This invention realizes a system that processes visual and audio data in real time to analyze the emotional state and behavioral patterns of viewers. It is implemented according to the following configuration.

[0511] The server collects and processes visual and audio data obtained through smart devices such as Oculus Quest in real time. This includes technologies for user eye tracking, facial recognition, and voice analysis. Specifically, it uses VisageSDK and OpenCV for facial expression analysis and Face API and EigenFace to identify microexpressions. This allows the server to identify the user's emotional state.

[0512] Furthermore, the servers process the collected data in a cloud environment. AWS Lambda or Azure ML is used to extract features from the data, and Scikit-learn is used to train machine learning models. These models are used to predict viewer behavior and contribute to ad optimization.

[0513] The device provides users with feedback to optimize advertising content based on the server's processing results. For example, if a smile is detected while a user is watching an advertisement for sports shoes, the data from that moment is highlighted and analyzed to clearly indicate which elements captured the user's attention.

[0514] For example, if a user shows a strong emotional reaction to a particular product or service while watching an advertisement, that information is used to improve marketing activities. This result is used by a generative AI model to create efficient prompts, presented in text format such as, "Tell us about your reaction to watching this shoe advertisement. What aspects particularly interested you or surprised you?"

[0515] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0516] Step 1:

[0517] Users view advertisements using smart devices. The devices collect visual and audio data in real time. Inputs include the user's gaze, facial expressions, and audio information, while outputs are processable digital data.

[0518] Step 2:

[0519] The server preprocesses the collected data using VisageSDK and OpenCV. In this step, it performs eye tracking and face recognition using the raw input data, converting the user's facial expressions and gaze direction into digital data. The output consists of facial expression data and gaze data.

[0520] Step 3:

[0521] The server utilizes Face API and EigenFace to analyze the user's emotional state from facial expression data. The input is facial expression data, and the output is data indicating the user's emotional state. Here, the server identifies whether the user's emotion is laughter, surprise, or interest.

[0522] Step 4:

[0523] Emotional state data and behavioral pattern data are integrated on a server, and feature extraction is performed. Using a cloud environment, the integrated input data is analyzed, and features indicating specific emotions and behavioral patterns related to advertising are extracted. The output is specific features designed to enhance advertising effectiveness.

[0524] Step 5:

[0525] The server uses Scikit-learn to train a machine learning model and predict future audience reactions. The input is feature data, and the output is the model's result predicting future audience reactions.

[0526] Step 6:

[0527] The device optimizes advertising content based on prediction results generated by the server. It displays ads that respond to the user's emotions and behavior to maximize effectiveness. Here, ad display is adjusted according to the user's specific emotions and actions.

[0528] Step 7:

[0529] The server uses a generation AI model to create prompt messages. The input is user response data, and the output is the prompt message provided to the user. The generated message will be in the form of text such as, "Please tell us about your reaction to watching this shoe advertisement. What aspects particularly interested you or surprised you?"

[0530] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0531] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0532] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[0533] [Fourth Embodiment]

[0534] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0535] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[0536] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0537] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[0538] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0539] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0540] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0541] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0542] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0543] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0544] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0545] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0546] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0547] This invention is a system for analyzing viewer activity data and optimizing marketing activities based on that data. This system is broadly composed of the following processes: data collection, data analysis, predictive model generation, and result feedback.

[0548] First, the server collects diverse data in real time from viewers of the product presentation video. It captures behavioral data such as viewing time, play / pause events, and skipped sections, and converts it into a format that can be immediately analyzed. This process prepares the foundational data necessary to understand the behavioral characteristics of the viewers.

[0549] Next, the server performs analysis based on the collected data. It extracts viewer behavior patterns as features and uses them to run a machine learning model to predict the likelihood of inquiries or applications. This makes it possible to numerically visualize how interested viewers are in the product. This model is designed to make highly accurate predictions by utilizing past training data.

[0550] The server then analyzes the prediction results and generates a strategic report in a format that the marketing team can use. This report includes content that resonated with viewers, responses by segment, and areas for improvement, and is provided to each user's device. This makes it easier for the marketing team to make decisions regarding adjusting video content and executing their next campaign strategies.

[0551] For example, when a user watches a video, sections where they frequently rewind or watch for extended periods are recorded. The server analyzes this data to determine which sections strongly capture the viewer's interest. This information is reflected in reports and used by the marketing team as valuable insights to create new promotional videos that highlight those sections.

[0552] The following describes the processing flow.

[0553] Step 1:

[0554] The server monitors the activity of users watching the product presentation videos in real time and collects data from viewers. Specifically, it acquires and records detailed event data such as the start time of viewing, the timing of playback and stopping, the sections skipped, and the viewing completion rate.

[0555] Step 2:

[0556] The server performs preprocessing on the collected raw data. Preprocessing includes organizing the data, interpolating missing values, and removing noisy data. This process creates a well-organized dataset that is suitable for subsequent analysis.

[0557] Step 3:

[0558] The server extracts viewer behavior patterns as features from the pre-processed data. In this step, multidimensional features such as average viewing time, rewind frequency, and skipping tendencies are calculated. This quantifies viewer interest and engagement.

[0559] Step 4:

[0560] The server uses a machine learning model based on extracted features to predict the likelihood of inquiries or applications. The machine learning model is pre-trained to achieve precise predictions. The model's output represents the probability of a response corresponding to each viewing session.

[0561] Step 5:

[0562] The server receives the prediction results and generates a visual strategic report. This report includes interesting points about viewing patterns and areas for improvement, providing concrete suggestions for marketing activities.

[0563] Step 6:

[0564] The terminal receives strategic reports provided by the server, which are then reviewed by members of the marketing team. This helps them optimize their video content editing and distribution strategies for the future.

[0565] (Example 1)

[0566] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0567] When building effective marketing strategies using viewer behavior data, it is essential to accurately understand viewer interests and efficiently optimize strategies. However, currently, the process from data collection to analysis and strategy proposal is fragmented, making it difficult to respond quickly and make highly accurate predictions.

[0568] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0569] In this invention, the server includes means for collecting and preprocessing viewer behavior information from terminals in real time, means for extracting feature quantities that indicate behavioral trends from the collected behavior information, and means for utilizing a generative AI model that predicts responses to inquiries and suggestions based on the extracted feature quantities. This enables a rapid and highly accurate understanding of viewer interests and the optimization of effective marketing strategies.

[0570] "Device" refers to an electronic device used by viewers to watch video, and includes personal computers, smartphones, tablets, and other similar devices.

[0571] "Viewers" refers to individuals or groups who watch video content and whose viewing behavior is collected as data.

[0572] "Action information" refers to data about the actions viewers perform while watching a video, and specifically includes viewing time, playback, pause, skip, and other events.

[0573] "Collecting data in real time" refers to the process of instantly sending viewer behavior information to the server without delay after an action occurs and acquiring it as data.

[0574] "Preprocessing" refers to the data manipulation required to prepare collected raw data for analysis, and includes imputing missing values, removing outliers, and converting to an appropriate data format.

[0575] In data analysis, "features" refer to data that quantifies patterns or statistical properties used by models for learning and prediction.

[0576] A "generative AI model" refers to an artificial intelligence model that is built to extract regularities and patterns from given data and to make predictions and classifications based on new data.

[0577] "Market research activities" refer to a series of analyses and strategy-building processes conducted to understand market trends for a particular product or service.

[0578] A "strategic report" is a document used to consider the next marketing steps based on the results of market research activities, and refers to a report that includes data analysis results and recommended measures.

[0579] This invention relates to a data analysis system for optimizing marketing activities by utilizing viewer behavior data. This system uses a device on which users view video content and collects their viewing behavior as data. The device consists of personal computers, smartphones, tablets, etc. A server collects viewer behavior information from these devices in real time. This behavior information includes viewing time, playback, pause, skip, and other events, and a JavaScript code snippet runs on the device to send the data to the server.

[0580] Subsequently, the server preprocesses the collected data, performing tasks such as imputing missing values ​​and removing outliers, and converting the data into a unified format using libraries such as Python's Pandas. From the preprocessed data, the server extracts features of viewer behavior patterns and prepares them as input data for a machine learning model. The generative AI model utilizes a pre-trained neural network using libraries such as Scikit-learn and TensorFlow.

[0581] The server uses a generative AI model to predict viewers' levels of interest, inquiries, and responses to suggestions. This makes it possible to quantify and visualize the degree to which viewers are interested in a product or service. Based on these predictions, the server generates a strategic report for market research activities and provides it to each remote terminal.

[0582] For example, if a user repeatedly plays a specific section while watching a product explanation video, the server will determine that the user has a high level of interest in that section and reflect this information in the report. The marketing team can then use this report to create new promotional content that highlights that section.

[0583] An example of a prompt would be: "Analyze the predictive report generated from viewing data and propose a new promotional strategy that highlights the sections that particularly captured audience interest."

[0584] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0585] Step 1:

[0586] The device collects real-time information about the user's actions while watching videos. This information includes viewing time, playback, pause, skipping, etc. A JavaScript code snippet is used to build a system that periodically sends this data to the server. The input to this process is the user's viewing behavior, and the output is the behavioral data sent to the server.

[0587] Step 2:

[0588] The server preprocesses the operational information received from the terminal. Specifically, it uses the Python Pandas library to impute missing values ​​and filter out outliers. Afterward, it converts the data into a unified format, preparing it for analysis. The input for this step is the raw data received from the terminal, and the output is the preprocessed, purified data.

[0589] Step 3:

[0590] The server extracts features from pre-processed data. It quantifies viewer behavior patterns, such as the distribution of viewing time and the proportion of frequently played sections. Statistical methods and data analysis techniques are used for this feature extraction. The input for this step is pre-processed data, and the output is a feature vector.

[0591] Step 4:

[0592] The server processes feature vectors using a generative AI model to predict viewer interest and likelihood of response. It then applies a pre-trained model using Scikit-learn or TensorFlow to calculate a prediction score. The input for this step is a feature vector, and the output is the prediction score.

[0593] Step 5:

[0594] The server generates a strategic report on market research activities based on the prediction score. Visualization tools are used to format the report in a visually easy-to-understand manner. The report includes content sections that viewers showed particular interest in, as well as recommended improvement actions. The input for this step is the prediction score, and the output is the strategic report.

[0595] Step 6:

[0596] The user (marketing team) uses the generated strategic report to optimize their next marketing strategy. Based on the feedback in the report, they implement specific measures, such as creating new promotional content. The input for this step is the strategic report, and the output is the improved marketing activities.

[0597] (Application Example 1)

[0598] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0599] Traditional methods of utilizing viewer data have made it difficult to efficiently analyze large amounts of data and formulate accurate marketing strategies based on that analysis. Furthermore, providing content that reflects individual viewer interests and behavioral patterns in real time has been challenging, resulting in difficulties in improving customer satisfaction and business performance.

[0600] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0601] In this invention, the server includes means for capturing and pre-processing viewer activity data in real time, means for extracting behavioral pattern features from the captured data, and means for tracking viewer responses to each viewing section in real time and providing personalized content suggestions. This enables direct utilization of viewer reactions to provide dynamic content tailored to customer interests and optimize marketing strategies.

[0602] "Activity data" refers to data about viewers' actions when viewing content, including viewing time and events such as playback, pausing, and skipping.

[0603] "Preprocessing" refers to the process of preparing data into an analyzable format, including tasks such as data cleaning and format conversion.

[0604] "Behavioral pattern features" are analytical indicators extracted from viewer behavior data, and are elements used to quantify viewing trends and interests.

[0605] "Personalized content recommendations" is an approach that presents specific content tailored to an individual's interests based on their behavioral data.

[0606] "Real-time tracking" refers to the process of instantly recording and retaining viewer actions as data the moment they occur.

[0607] In the system implementing this invention, the server captures viewer activity data in real time and preprocesses the data to prepare it for analysis. The server then uses machine learning to extract behavioral pattern features from the captured data and evaluates the viewer's level of interest. Based on this, personalized content suggestions are made in real time.

[0608] The hardware includes smartphones and tablet devices used by users, which have means of communication to send viewing data to the server. The software uses programs such as Python and TensorFlow for data preprocessing and analysis using machine learning models.

[0609] As a concrete example, when a user watches a particular promotional video, actions such as playing, pausing, and rewinding are recorded. The server analyzes this viewing data to identify sections that the user has watched repeatedly and provides the user with additional content related to those sections. This allows users to delve deeper into the topics that interest them.

[0610] An example of a prompt to a generative AI model would be, "Analyze the content of the most viewed sections and generate additional content to recommend to the user." Using this prompt, the AI ​​model can demonstrate its ability to create valuable content based on viewing data.

[0611] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0612] Step 1:

[0613] The device captures user viewing data. It records viewing time, playback, pause, skip, and other events in real time and sends this data to the server. The input is the user's action events, and the output is the viewing event data sent to the server.

[0614] Step 2:

[0615] The server receives and preprocesses the viewing data. It performs data cleaning and converts it into a format that can be processed by machine learning models. The input is viewing event data from the terminal, and the output is formatted viewing data.

[0616] Step 3:

[0617] The server extracts behavioral pattern features from formatted data. It extracts statistical features from viewing data and processes them to quantify interest trends. The input is pre-processed viewing data, and the output is the features of the viewers' behavioral patterns.

[0618] Step 4:

[0619] Based on the features extracted by the server, a generative AI model is used to predict the user's level of interest and appropriate content. Prompt messages are input to the AI ​​model, and prediction results are generated. The input consists of behavioral pattern features and prompt messages, while the output is the level of interest and recommended content.

[0620] Step 5:

[0621] The server provides personalized content to the device based on prediction results. It takes the user's viewing history into consideration and prioritizes suggesting content that might be of interest. The input is the prediction result, and the output is a list of recommended content for the device.

[0622] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0623] This invention is a system that analyzes viewer viewing activity data and emotions in real time to help optimize marketing activities. The system includes functions for data collection, emotion recognition, data analysis, predictive model generation, and result feedback.

[0624] First, the server collects viewing activity data in real time from users watching the product presentation. This includes data such as viewing time, playback / pause actions, and skipped sections. In parallel, the user's video and audio data is fed into the emotion engine for emotion recognition. The emotion engine identifies emotional states such as joy, surprise, and dissatisfaction through facial expression analysis, voice tone analysis, and other methods.

[0625] Next, the server integrates the collected viewing activity data with user sentiment data and extracts features. This allows for the quantification of not only viewers' interests and concerns, but also their intuitive responses based on emotions. This data is then fed into a machine learning model to predict the likelihood of inquiries and applications with high accuracy.

[0626] The server then receives the results estimated by the machine learning model and generates a strategic report for improving marketing activities. This report visualizes changes in viewer sentiment and the resulting distribution of interests, and is provided to the device. This allows marketers to develop audience-optimized strategies from the perspective of video content.

[0627] For example, if a user experiences surprise while watching a video explaining a specific feature of a product, the server will specifically highlight and analyze the data from that moment. The marketing team can then implement a PDCA cycle to stimulate potential purchase intent by reinforcing the elements that evoked surprise and expanding upon them in subsequent videos.

[0628] The following describes the processing flow.

[0629] Step 1:

[0630] The server collects real-time activity data from viewers (users) during the product presentation video, including viewing time, playback, pause, and skipping. Simultaneously, it captures video and audio data in real time via cameras and microphones and supplies this data to the emotion engine.

[0631] Step 2:

[0632] The server uses an emotion engine to recognize the user's emotions from the captured video and audio data. The emotion engine analyzes changes in facial expressions and tone of voice to identify emotions such as joy, surprise, dissatisfaction, and boredom. This allows the user's emotional progression during viewing to be stored as data.

[0633] Step 3:

[0634] The server integrates viewing activity data and emotion recognition data to extract features. These features include high interest in specific video segments, frequency of emotion changes, and viewing time distribution for each type of emotion. This allows for a comprehensive understanding of user behavior patterns and emotional responses.

[0635] Step 4:

[0636] The server uses a machine learning model based on extracted features to predict the likelihood of inquiries or applications. The trained model performs inferences based on historical data and calculates the probability of each user responding.

[0637] Step 5:

[0638] The server analyzes the prediction results and generates a visually organized strategic report. This report includes marketing suggestions, such as the points that viewers were most interested in and the moments when their emotions shifted significantly.

[0639] Step 6:

[0640] The terminal (used by the marketing team) receives reports from the server and uses them to create marketing strategies based on viewer sentiment and behavior data. This helps optimize future video content and design campaigns.

[0641] (Example 2)

[0642] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0643] Conventional audience data analysis systems have difficulty understanding viewers' emotional states, and marketing strategies that take viewers' intuitive reactions into account have not been adequately developed. As a result, content optimization and the accurate capture of viewers' interests have been hindered, limiting the effectiveness of marketing activities.

[0644] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0645] This invention includes a server that captures and preprocesses viewer activity and emotional data in real time, extracts behavioral patterns and emotional state features from the captured data, and utilizes a machine learning model that predicts responses to inquiries and applications based on the extracted features. This enables the optimization of content that reflects the emotional state of viewers and the formulation of highly accurate marketing strategies.

[0646] "Viewer activity data" refers to information about viewers' behavior when viewing content, including viewing time, playback, pausing, skipping, and other events.

[0647] "Emotional data" refers to information that represents the emotional state of the viewer, and includes data that identifies emotional states such as joy, surprise, and dissatisfaction based on facial expressions and tone of voice.

[0648] "Means of preprocessing" refers to a mechanism that processes viewer activity data and emotional data to prepare them for analysis.

[0649] A "method for extracting features" is a mechanism for identifying important patterns and characteristics from the data being analyzed and expressing them numerically or otherwise.

[0650] A "machine learning model" is an algorithm or mathematical model that learns patterns and relationships from data to predict future trends and outcomes.

[0651] A "strategic report" is a strategic document generated based on collected and analyzed data, intended to improve marketing activities.

[0652] "Visualization" refers to depicting and representing complex data and information in an easily understandable way, primarily using graphs and charts.

[0653] This system aims to optimize marketing strategies by collecting and analyzing viewer activity and emotional data in real time. The server collects activity data such as viewing time, playback, pause, and skipping while users are watching product presentation videos. In addition, it uses input devices such as cameras and microphones to capture the user's facial expressions and voice in real time for the emotional analysis engine. Emotional analysis includes facial expression analysis and voice tone analysis, using common analysis software to identify emotional states.

[0654] The server integrates this collected data and extracts features using a programming language like Python and the pandas library. In this process, viewer behavior and emotional patterns are represented as numerical data and formatted into a format that can be input into machine learning frameworks such as TensorFlow and Scikit-learn. Next, the server uses a pre-trained machine learning model based on these features to predict the likelihood of inquiries and applications. The machine learning model analyzes viewer behavior and emotional trends based on previously accumulated data and predicts future behavior with high accuracy.

[0655] Furthermore, the server automatically generates strategic reports utilizing the prediction results of machine learning models. These reports visually represent changes in viewer emotions and viewing data, and are provided to the terminal, making them easily understandable for marketing personnel. This allows them to quickly formulate optimal marketing strategies tailored to the emotional state of the viewers.

[0656] For example, a user might react with surprise when a specific feature of a product is explained to them. In this case, the server highlights the feature data from that moment, revealing key areas for improvement for the marketing team. By emphasizing the elements that evoked this surprise in the next video, the viewer's interest can be further stimulated.

[0657] An example of a prompt is, "How can we identify specific moments that surprised viewers during a product presentation?" This allows for the strategic improvement of video content by leveraging viewer emotions.

[0658] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0659] Step 1:

[0660] The user watches a product presentation video. The server collects user viewing activity data during this process. Data such as viewing time, playback, pause, and skipping are obtained as input. This input data is stored on the server and structured in preparation for analysis.

[0661] Step 2:

[0662] The server collects the user's facial and voice data in real time and feeds it into the emotion recognition engine. Specifically, it uses data obtained from the camera and microphone as input to identify emotional states such as joy, surprise, and dissatisfaction using facial expression analysis and voice tone analysis. As a result of this processing, data on the emotional state is output.

[0663] Step 3:

[0664] The server integrates viewing activity data and sentiment data. Using the Python pandas library, it extracts features from both datasets. Feature extraction yields data quantifying changes in viewers' interest, attention, and sentiment. This data is used as input for machine learning processing.

[0665] Step 4:

[0666] The server inputs feature data into TensorFlow or Scikit-learn and analyzes it using a pre-trained machine learning model. The input is feature data, and the output generates predictions for inquiries and applications. These prediction results are then used to develop marketing strategies.

[0667] Step 5:

[0668] The server generates a strategic report based on the forecast results. The report includes changes in audience sentiment and the resulting trends in viewing activity. The report is sent to terminals and made accessible to marketers. This allows for the development of optimized content strategies for audiences based on visualized data.

[0669] (Application Example 2)

[0670] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0671] Traditional advertising suffers from insufficient detailed analysis of viewers' emotions and behavior, making it difficult to formulate effective marketing strategies. To solve this problem, a method is needed that optimizes advertising using real-time data obtained from viewers' visual and auditory experiences.

[0672] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0673] In this invention, the server includes means for collecting and processing visual and audio data in real time, means for deriving behavioral patterns and emotional states from the collected data, and means for utilizing machine learning models to predict responses based on the derived features and states. This enables the optimization of advertisements based on the emotions and behavior of viewers.

[0674] "Visual data" refers to digital information that captures visual information such as viewers' gaze, facial expressions, and actions in real time.

[0675] "Audio data" refers to digital information that captures auditory information, such as the listener's voice and tone of voice.

[0676] "Means for collecting and processing data in real time" refers to technologies that perform the process of instantly acquiring digital data and converting it into an analyzable format.

[0677] "Behavioral patterns" refer to the regularity of the visual and behavioral behaviors of viewers when they view advertisements.

[0678] "Emotional state" refers to the psychological response of viewers, identifying emotions such as joy, anger, sadness, and happiness.

[0679] A "machine learning model" is an algorithm that learns specific patterns from large amounts of data and uses them to predict future reactions and behaviors.

[0680] "Ad optimization" is the process of adjusting the content and presentation of advertisements based on the interests and concerns of the audience in order to maximize their effectiveness.

[0681] This invention realizes a system that processes visual and audio data in real time to analyze the emotional state and behavioral patterns of viewers. It is implemented according to the following configuration.

[0682] The server collects and processes visual and audio data obtained through smart devices such as Oculus Quest in real time. This includes technologies for user eye tracking, facial recognition, and voice analysis. Specifically, it uses VisageSDK and OpenCV for facial expression analysis and Face API and EigenFace to identify microexpressions. This allows the server to identify the user's emotional state.

[0683] Furthermore, the servers process the collected data in a cloud environment. AWS Lambda or Azure ML is used to extract features from the data, and Scikit-learn is used to train machine learning models. These models are used to predict viewer behavior and contribute to ad optimization.

[0684] The device provides users with feedback to optimize advertising content based on the server's processing results. For example, if a smile is detected while a user is watching an advertisement for sports shoes, the data from that moment is highlighted and analyzed to clearly indicate which elements captured the user's attention.

[0685] For example, if a user shows a strong emotional reaction to a particular product or service while watching an advertisement, that information is used to improve marketing activities. This result is used by a generative AI model to create efficient prompts, presented in text format such as, "Tell us about your reaction to watching this shoe advertisement. What aspects particularly interested you or surprised you?"

[0686] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0687] Step 1:

[0688] Users view advertisements using smart devices. The devices collect visual and audio data in real time. Inputs include the user's gaze, facial expressions, and audio information, while outputs are processable digital data.

[0689] Step 2:

[0690] The server preprocesses the collected data using VisageSDK and OpenCV. In this step, it performs eye tracking and face recognition using the raw input data, converting the user's facial expressions and gaze direction into digital data. The output consists of facial expression data and gaze data.

[0691] Step 3:

[0692] The server utilizes Face API and EigenFace to analyze the user's emotional state from facial expression data. The input is facial expression data, and the output is data indicating the user's emotional state. Here, the server identifies whether the user's emotion is laughter, surprise, or interest.

[0693] Step 4:

[0694] Emotional state data and behavioral pattern data are integrated on a server, and feature extraction is performed. Using a cloud environment, the integrated input data is analyzed, and features indicating specific emotions and behavioral patterns related to advertising are extracted. The output is specific features designed to enhance advertising effectiveness.

[0695] Step 5:

[0696] The server uses Scikit-learn to train a machine learning model and predict future audience reactions. The input is feature data, and the output is the model's result predicting future audience reactions.

[0697] Step 6:

[0698] The device optimizes advertising content based on prediction results generated by the server. It displays ads that respond to the user's emotions and behavior to maximize effectiveness. Here, ad display is adjusted according to the user's specific emotions and actions.

[0699] Step 7:

[0700] The server uses a generation AI model to create prompt messages. The input is user response data, and the output is the prompt message provided to the user. The generated message will be in the form of text such as, "Please tell us about your reaction to watching this shoe advertisement. What aspects particularly interested you or surprised you?"

[0701] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0702] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0703] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[0704] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[0705] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[0706] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[0707] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[0708] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[0709] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[0710] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[0711] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[0712] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[0713] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[0714] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[0715] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[0716] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[0717] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[0718] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[0719] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[0720] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[0721] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[0722] The following is further disclosed regarding the embodiments described above.

[0723] (Claim 1)

[0724] A means for capturing and pre-processing viewer activity data in real time,

[0725] A means of extracting behavioral pattern features from captured data,

[0726] A method that utilizes machine learning models to predict responses to inquiries and applications based on extracted features,

[0727] A means of generating and proposing strategic reports on marketing activities based on prediction results,

[0728] A system that includes this.

[0729] (Claim 2)

[0730] The system according to claim 1, comprising means for generating identification information using event data such as viewing time, playback, pause, and skip, which indicate the viewer's interest.

[0731] (Claim 3)

[0732] The system according to claim 1, further comprising means for providing a report generated on a remote terminal and providing feedback to optimize video content based on viewer behavior data.

[0733] "Example 1"

[0734] (Claim 1)

[0735] A means for collecting viewer behavior information from a terminal in real time and performing preprocessing,

[0736] A means for extracting features that indicate behavioral tendencies from collected motion information,

[0737] A method that utilizes a generative AI model to predict responses to inquiries and suggestions based on extracted features,

[0738] A means of generating and proposing strategic reports for market research activities based on forecast results,

[0739] A data analysis system that includes this.

[0740] (Claim 2)

[0741] The data analysis system according to claim 1, comprising means for generating identification information using viewing time and event information such as playback, pause, and skip that indicate the viewer's interest.

[0742] (Claim 3)

[0743] The data analysis system according to claim 1, comprising means for providing a report generated on a remote terminal and for providing improvement measures to optimize video material based on viewer behavior information.

[0744] "Application Example 1"

[0745] (Claim 1)

[0746] A means for capturing and pre-processing viewer activity data in real time,

[0747] A means of extracting behavioral pattern features from captured data,

[0748] A method that utilizes machine learning models to predict responses to inquiries and applications based on extracted features,

[0749] A means of generating and proposing strategic reports on marketing activities based on prediction results,

[0750] A means of tracking viewer reactions to each viewing section in real time and providing personalized content recommendations,

[0751] A system that includes this.

[0752] (Claim 2)

[0753] A means for generating identification information using event data such as viewing time, playback, pause, and skip, which indicate the viewer's interest,

[0754] The system according to claim 1, comprising means for dynamically adjusting video content based on viewer behavior.

[0755] (Claim 3)

[0756] A means of providing reports generated on remote terminals and providing feedback to optimize video content based on viewer behavior data,

[0757] The system according to claim 1, comprising means for providing promotional content based on sections that attract the interest of viewers.

[0758] "Example 2 of combining an emotion engine"

[0759] (Claim 1)

[0760] A means for capturing and pre-processing viewer activity data and emotional data in real time,

[0761] A means for extracting behavioral patterns and emotional state features from captured data,

[0762] A method that utilizes machine learning models to predict responses to inquiries and applications based on extracted features,

[0763] A means of generating and proposing strategic reports on marketing activities based on prediction results,

[0764] A means of providing feedback to optimize content based on viewer behavior and sentiment data by providing reports to remote terminals,

[0765] A system that includes this.

[0766] (Claim 2)

[0767] The system according to claim 1, comprising means for generating identification information using viewing event data indicating viewer interest and emotion data indicating emotional state.

[0768] (Claim 3)

[0769] The system according to claim 1, comprising means for formulating a marketing strategy based on the interests and emotions of viewers by utilizing the generated reports.

[0770] "Application example 2 when combining with an emotional engine"

[0771] (Claim 1)

[0772] A means for collecting and processing visual and audio data in real time,

[0773] A means of deriving behavioral patterns and emotional states from collected data,

[0774] A means of using a machine learning model to predict a response based on the derived features and state,

[0775] A means of presenting information for optimizing advertising activities based on predicted responses,

[0776] A system that includes this.

[0777] (Claim 2)

[0778] The system according to claim 1, comprising means for generating individual identification information using visual data, audio data, and behavioral information.

[0779] (Claim 3)

[0780] The system according to claim 1, further comprising means for supplying information to a remote device and providing feedback to assist in the optimization of advertising content. [Explanation of Symbols]

[0781] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A means for capturing and pre-processing viewer activity data in real time, A means of extracting behavioral pattern features from captured data, A method that utilizes machine learning models to predict responses to inquiries and applications based on extracted features, A means of generating and proposing strategic reports on marketing activities based on prediction results, A system that includes this.

2. The system according to claim 1, comprising means for generating identification information using event data such as viewing time, playback, pause, and skip, which indicate the viewer's interest.

3. The system according to claim 1, further comprising means for providing a report generated on a remote terminal and providing feedback to optimize video content based on viewer behavior data.

Images