Content evaluation system, content evaluation method, program, and server device

Abstract

To provide a content evaluation system, a content evaluation method, a program, and a server device for outputting evaluation of a user who has viewed a content with high accuracy.SOLUTION: In a content evaluation system 1, an application server device includes: an acquisition unit which acquires a content from a user terminal device; and an evaluation unit which evaluates the content on the basis of a biometric signal calculated by a calculation unit. An AI server device includes a receiving unit which receives the content, and a prediction unit which inputs the content to a machine learning model to predict the biometric signal on the basis of an output of the machine learning model. The application server device functions as the calculation unit which calculates the biometric signal of a user who has viewed the content. The machine learning model is configured to learn parameters so as to output sensor data when an experimental content is input, while using biometric signals as sensor data acquired from a sensor when the user has viewed the experimental content as learning data.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a content evaluation system, a content evaluation method, a program, and a server device. [Background technology]

[0002] Conventionally, an advertisement evaluation system for evaluating the effectiveness of an advertisement using an advertisement display device is known, for example from Patent Document 1. This advertisement evaluation system acquires behavioral information of people who pass near the advertisement display device and those people who view the advertisement, and evaluates the effectiveness of the advertisement based on the behavioral information after the advertisement is displayed by the advertisement display device. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2023-66306 A Summary of the Invention [Problem to be solved by the invention]

[0004] The advertising evaluation system described in Patent Document 1 evaluates the effectiveness of an advertisement based on the actions of people who see the advertisement, but there are reasons other than the advertisement that cause people to take action, and it may not be possible to evaluate the effectiveness of the advertisement with high accuracy.

[0005] The present disclosure has been made in consideration of the above circumstances, and aims to provide a content evaluation system, a content evaluation method, a program, and a server device that can output evaluations by users who have viewed content with high accuracy. [Means for solving the problem]

[0006] The present disclosure has been made to solve the above-mentioned problems, and one aspect of the present disclosure is a content evaluation system including an acquisition unit that acquires content, a calculation unit that calculates a biosignal of a user who has viewed the content, and an evaluation unit that evaluates the content based on the biosignal calculated by the calculation unit.

[0007] Another aspect of the present disclosure is a content evaluation method including the steps of: an information processing device acquiring content; the information processing device calculating a biosignal of a user who has viewed the content; and the information processing device evaluating the content based on the calculated biosignal.

[0008] Another aspect of the present disclosure is a program that causes a computer of an information processing device to execute the steps of acquiring content, calculating a biosignal of a user who has viewed the content, and evaluating the content based on the calculated biosignal.

[0009] Another aspect of the present disclosure is a server device comprising a reception unit that receives content, and a prediction unit that inputs the content into a machine learning model and predicts a biological signal based on an output of the machine learning model, wherein the machine learning model uses as learning data a biological signal as sensor data obtained from a sensor when a user is made to view experimental content, and parameters are learned to output sensor data when the experimental content is input. Effect of the Invention

[0010] According to one aspect of the present invention, it is possible to output evaluations of users who have viewed content with high accuracy. [Brief description of the drawings]

[0011] [Figure 1] 1 is a block diagram showing an example of a configuration of a content evaluation system according to an embodiment. [Diagram 2]FIG. 2 is a block diagram showing an example of an evaluation unit in the embodiment. [Diagram 3] FIG. 2 is a block diagram showing an example of a learning data generation process according to an embodiment. [Figure 4] FIG. 2 is a block diagram showing an example of a learning process according to the embodiment. [Diagram 5] FIG. 4 is a diagram showing an example of a content selection screen in the embodiment. [Figure 6] FIG. 13 is a diagram showing an example of an attribute selection screen in the embodiment. [Figure 7] FIG. 13 is a diagram showing an example of a diagnosis result screen in the embodiment. [Figure 8] FIG. 13 is a diagram illustrating an example of a heat map according to the embodiment. [Figure 9] FIG. 4 is a diagram showing an example of a group ID, a title, and a result ID in the embodiment. [Figure 10] FIG. 11 is a diagram showing an example of a diagnosis result of a plurality of contents in the embodiment. [Figure 11] 13A and 13B are diagrams showing examples of screens for comparing diagnostic results by changing attributes in the embodiment. [Figure 12] FIG. 13 is a diagram showing an example of a screen for comparing a plurality of pieces of content in the embodiment. [Figure 13] FIG. 13 is a diagram showing an example of a screen for comparing pre-edited content with post-edited content in the embodiment. [Figure 14] 10 is a sequence diagram showing an example of a processing procedure of a content evaluation system according to an embodiment. FIG. [Figure 15] FIG. 11 is a sequence diagram showing another example of the processing procedure of the content evaluation system in the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] A content evaluation system, a content evaluation method, a program, and a server device to which the present invention is applied will be described below with reference to the drawings.

[0013] (Configuration of Content Evaluation System 1) FIG. 1 is a block diagram showing an example of a configuration of a content evaluation system 1 in an embodiment. The content evaluation system 1 includes, for example, an application server device 100, an AI (artificial intelligence) server device 200, and a user terminal device 300. The application server device 100, the AI ​​server device 200, and the user terminal device 300 may have a communication interface such as a NIC (Network Interface Card) or a wireless communication module for connecting to a network such as the Internet, and may exchange information with each other. The network may include, for example, a general-purpose network such as the Internet, and a private network such as local 5G or WiFi (registered trademark).

[0014] The application server device 100 is an information processing device that performs various processes by, for example, having a CPU (Central Processing Unit) execute a program. The application server device 100 includes, for example, an acquisition unit 102 and an evaluation unit 106. The acquisition unit 102 and the evaluation unit 106 are functional units that are realized by the CPU executing a program. In addition, some or all of these functional units may be realized by hardware (including circuitry) such as an LSI (Large Scale Integration), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a GPU (Graphics Processing Unit), or may be realized by a combination of software and hardware.

[0015] The acquisition unit 102 acquires content. The evaluation unit 106 evaluates the content based on the biosignal calculated by the AI ​​server device 200. The AI ​​server device 200 calculates the biosignal of the user who viewed the content. Note that the acquisition unit 102 and the evaluation unit 106 are provided in a single information processing device, but are not limited to this, and may be provided in one or more information processing devices other than the application server device 100, or may be built into the user terminal device 300.

[0016] The AI ​​server device 200 is, for example, an information processing device that performs various processes by having a CPU execute a program. The AI ​​server device 200 functions as a calculation unit that calculates the biosignals of a user who has viewed the content. The AI ​​server device 200 includes, for example, a reception unit 202, a prediction unit 204, a learning processing unit 206, a learning data storage unit 208, and an algorithm storage unit 210. The reception unit 202, the prediction unit 204, and the learning processing unit 206 are functional units that are realized by the CPU executing a program. The learning data storage unit 208 and the algorithm storage unit 210 are realized, for example, by a random access memory (RAM), a read only memory (ROM), a hard disk drive (HDD), a flash memory, or a hybrid storage device that combines a plurality of these. The training data storage unit 208 and the algorithm storage unit 210 may be realized by a single device, or some or all of the training data storage unit 208 and the algorithm storage unit 210 may be external devices accessible by the AI ​​server device 200, such as a NAS or an external storage server.

[0017] The reception unit 202 receives the content to be evaluated from the application server device 100. The prediction unit 204 predicts the biosignal based on the content received by the reception unit 202. The prediction unit 204 may predict a subjective evaluation when the content is viewed as the biosignal. The subjective evaluation is, for example, the likeability of the content felt by the user and the readability of the content. The prediction unit 204 predicts the biosignal using a predetermined algorithm. The prediction unit 204 outputs the predicted biosignal to the calculation unit 104. The predetermined algorithm is, for example, a machine learning model. The machine learning model is, for example, a deep learning model using a CNN (Convolutional Neural Network) or the like, but is not limited thereto, and a model applied to existing machine learning, such as a DCNN (DeepCNN), a decision tree, a hierarchical Bayes, or an SVM (Support Vector Machine), may be used. The learning processing unit 206 performs a process of learning the predetermined algorithm. The learning data storage unit 208 stores learning data for learning the predetermined algorithm. The algorithm storage unit 210 stores data that identifies a given algorithm.

[0018] In the embodiment, the AI ​​server device 200 is provided with a function for predicting a biological signal based on content, but the application server device 100 may be provided with a function for predicting a biological signal.

[0019] The user terminal device 300 is an information processing device such as a personal computer operated by a user. The user terminal device 300 includes, for example, a user interface unit 302, an application processing unit 304, and a content storage unit 306. The user interface unit 302 includes, for example, a display device, and input devices such as a keyboard and a mouse. The application processing unit 304 performs processing of a browser and application software, thereby causing the application server device 100 and the AI ​​server device 200 to evaluate the content. For example, the application processing unit 304 causes various information for evaluating the content to be displayed by the application processing unit 304, and transmits various requests for evaluating the content to the application server device 100 based on the user's operation.

[0020] 2 is a block diagram showing an example of the evaluation unit 106 in the embodiment. The biological signal is a signal indicating at least one change among cerebral blood flow, pupil diameter, gaze, skin potential, heart rate, subjective evaluation, and AB test when a user views content. The evaluation unit 106 calculates at least one score of interest, memory, attention level, likeability, and readability for the content.

[0021] Interest in content is an index that quantifies brain activity in the area of ​​interest in the prefrontal cortex, and indicates whether the content is personal, i.e., whether the content is attractive and worthy of attention. Memory for content is an index that measures the area of ​​working memory in the prefrontal cortex, and indicates whether the content stimulates thinking, i.e., how easily it is remembered. Attention to content is an index that scores the most frequently viewed areas of the 16 x 16 areas of the content, based on the product of pupil diameter and gaze retention time. Content preference is an index that uses an AB test model to construct an index of the probability that biological reactions are judged as favorable as input, and includes only positive factors for "interest," which includes positive and negative factors. Readability of content is an index of the probability that it is subjectively judged to be "easy to read."

[0022] The evaluation unit 106 includes, for example, a feature extraction unit 110, an interest evaluation model calculation unit 112, a memory evaluation model calculation unit 114, an attention evaluation model calculation unit 116, a desirability evaluation model calculation unit 118, a readability evaluation model calculation unit 120, a page ratio extraction model calculation unit 122, a desirability contribution heat map creation unit 124, and a gaze estimation heat map creation unit 126.

[0023] The feature extraction unit 110 extracts image features of the content acquired by the acquisition unit 102. The interest evaluation model calculation unit 112 calculates an interest score based on the content. The memory evaluation model calculation unit 114 inputs the biosignal calculated by the AI ​​server device 200 to a memory evaluation model, and calculates a memory score for the content using the output of the memory evaluation model. The attention evaluation model calculation unit 116 calculates an attention score for the content based on a gaze estimation heat map. The desirability evaluation model calculation unit 118 inputs the biosignal calculated by the AI ​​server device 200 to a desirability evaluation model, and calculates a desirability score for the content using the output of the desirability evaluation model. The readability evaluation model calculation unit 120 calculates a readability score for the content based on the page ratio.

[0024] The page ratio extraction model calculation unit 122 inputs the content to the page ratio evaluation model, and calculates the page ratio [%] for the content using the output of the page ratio evaluation model.

[0025] The desirability contribution heat map creation unit 124 creates a heat map showing the parts of the content that contributed to the desirability score, based on the content desirability score calculated by the desirability evaluation model calculation unit 118. The gaze estimation heat map creation unit 126 creates a heat map showing the gaze dwell rate within the content, based on the biological signal calculated by the AI ​​server device 200.

[0026] FIG. 3 is a block diagram illustrating an example of a process for generating learning data according to the embodiment. The learning data is generated by a learning data generation system. The learning data generation system includes, for example, an experimental content storage unit 400, a content playback device 402, and a sensor system 404. Experimental content is stored in the experimental content storage unit 400. The experimental content is content to be viewed by a subject, and is content in the same field as the content acquired by the application server device 100. The experimental content is, for example, content published in pamphlets, direct mail (DM), posters, magazine advertisements, booklets, etc. The experimental content may be, for example, content targeted at a specific attribute, or content targeted at a wide range of attributes. The content playback device 402 plays the experimental content when the subject is to view the content, and displays it on a display device.

[0027] The sensor system 404 includes, for example, a brain activity sensor, an eye tracker, a pupil response measuring and analyzing device, a skin potential sensor, a heart rate sensor, and an operation unit. The brain activity sensor detects biological signals such as cerebral blood flow signals. The eye tracker detects the position and dwell time of the gaze. The pupil response measuring and analyzing device detects the pupil diameter and the amount of change in the pupil diameter. The skin potential sensor collects instantaneous changes in the amount of skin sweating. The heart rate sensor extracts and quantifies the high-frequency components of the fluctuations in the heart rate. The operation unit obtains subjective evaluation data using the VAS method, for example, immediately after browsing the content. The operation unit allows the user to select a preferred content from two similar contents, and detects the answer result and the answer speed. The operation unit allows the user to select an easy-to-read content from two similar contents, and detects the answer result and the answer speed. The learning data storage unit 208 stores the content being played in association with the biological signal.

[0028] FIG. 4 is a block diagram showing an example of the learning process according to the embodiment. The AI ​​server device 200 includes a prediction unit 204 and a learning processing unit 206. The prediction unit 204 inputs the feature amount of the content to a machine learning model 204A, and outputs a predicted value of a biological signal for the content. The machine learning model 204A is, for example, a convolutional neural network (CNN).

[0029] The learning processing unit 206 constructs the machine learning model 204A. The machine learning model 204A acquires learning data from the learning data storage unit 208. The learning processing unit 206 inputs content to the machine learning model 204A during learning. The machine learning model 204A outputs a prediction result of a biological signal when the content is input. The learning processing unit 206 recursively updates the processing parameters of the machine learning model 204A so that the prediction result of the biological signal output from the machine learning model 204A matches the biological signal as learning data. The processing parameters are, for example, at least one of the number of layers, the number of nodes in each layer, the connection method of nodes between each layer, the activation function, the error function, and the gradient descent algorithm, the pooling region, the kernel, the weighting coefficient, and the weighting matrix in the convolutional neural network. As a result, the machine learning model 204A learns parameters so that the biological signal as sensor data acquired from the sensor system 404 when the user is made to watch the experimental content is used as learning data, and the sensor data is output when the experimental content is input. The learning data may include data indicating a subjective evaluation based on the user's operation in addition to biological signals detectable by the sensor.

[0030] The prediction unit 204 may input content to the machine learning model 204A corresponding to the attributes of the user, and calculate the biological signal based on the output of the machine learning model 204A. The machine learning model 204A uses the biological signal as sensor data acquired from the sensor system 404 when a user having the attributes is made to view experimental content as learning data, and learns parameters so as to output the sensor data when the experimental content is input.

[0031] The prediction unit 204 may input the content to a machine learning model corresponding to the medium of use in which the content is used, and calculate the biological signal based on the output of the machine learning model. The machine learning model 204A uses the biological signal as sensor data acquired from the sensor system 404 when the user is made to view the experimental content used in the medium of use as learning data, and learns parameters so as to output the sensor data when the experimental content is input.

[0032] The prediction unit 204 may input the content to a machine learning model 204A corresponding to the purpose of use for which the content is used, and calculate a biological signal based on the output of the machine learning model. The machine learning model 204A uses as learning data a biological signal as sensor data acquired from the sensor system 404 when the user is made to view experimental content used for the purpose of use, and learns parameters so as to output the sensor data when the experimental content is input.

[0033] FIG. 5 is a diagram showing an example of a content selection screen according to the embodiment. When the user accesses and logs in to the application server device 100, the user interface unit 302 displays a content selection screen as shown in Fig. 5. The content selection screen is a screen that displays content that was referenced in a past diagnosis. The user interface unit 302 uploads content (images) in response to the "New Diagnosis" button being selected, and displays an image information selection screen in response to the "Next" button being selected.

[0034] FIG. 6 is a diagram showing an example of an attribute selection screen according to the embodiment. In response to uploading the content to be diagnosed, the user interface unit 302 displays the attribute selection screen shown in FIG. 6. The attribute selection screen includes a title input field, a medium to be used selection field, and a purpose of use selection field. The medium to be used selection field includes check boxes for selecting pamphlet, magazine advertisement, poster, direct mail, or other. The purpose of use selection field includes check boxes for selecting impact / impression or information. In response to the "Diagnose" button being selected, the user interface unit 302 transmits a diagnosis request including the title, medium to be used, and purpose of use to the application server device 100.

[0035] FIG. 7 is a diagram showing an example of a diagnosis result screen according to the embodiment. The user interface unit 302 displays the diagnosis result screen shown in FIG. 7 based on the diagnosis result received from the application server device 100. The result diagnosis screen includes the title, the attribute, the content, and the score distribution of the diagnosis result. The result diagnosis screen may display the comment posted about the evaluation of the content. The user interface unit 302 may display the definition information of each of the interest score, the memory score, the attention score, the likeability score, and the readability score. For example, the user interface unit 302 may display the text indicating the definition of the interest score in response to the selection of the text display position of the interest score. The user interface unit 302 displays the diagnosis result corresponding to the attribute in response to the selection of the attribute. When multiple attributes are selected, the user interface unit 302 displays the result of multiplying the diagnosis results of each attribute.

[0036] The application processing unit 304 may display a total score of the content. The total score may be a rank such as S, A, or B based on the total value of the interest score, memory score, attention score, desirability score, and readability score. The total score may be a total score for each attribute. The total score for men may be, for example, the total value of the interest score, memory score, attention score, desirability score, and readability score for all age groups of men.

[0037] FIG. 8 is a diagram illustrating an example of a heat map according to the embodiment. The application processing unit 304 displays a heat map based on the interest score, memory score, attention score, favorability score, and readability score received from the application server device 100. The application processing unit 304 displays the page ratio [%] of the content based on the page ratio of the content calculated by the page ratio extraction model calculation unit 122. The page ratio may be displayed as the page ratio for each part into which the content is divided. The favorability contribution heat map creation unit 124 creates a heat map indicating the contribution to the favorability score for each part into which the content is divided. The gaze estimation heat map creation unit 126 creates a heat map based on the gaze dwell time for each part into which the content is divided. The application server device 100 can cause the application processing unit 304 to display the heat map by transmitting information of the created heat map to the user terminal device 300.

[0038] FIG. 9 is a diagram showing an example of a group ID, a title, and a result ID in the embodiment, and FIG. 10 is a diagram showing an example of a diagnosis result of a plurality of contents in the embodiment. The user interface unit 302 may set a group ID, a group title, and a result ID as shown in FIG. 9, and compare and display the diagnosis results of a plurality of contents as shown in FIG. 10. The diagnosis result includes, for example, the title, the medium of use, the purpose of use, and scores for each attribute such as a total score, a total desirability score, a male desirability score, etc., for each of the contents with image IDs "A", "B", and "C". This allows the user to compare the scores of a plurality of contents.

[0039] FIG. 11 is a diagram showing an example of a screen for comparing diagnostic results by changing attributes in the embodiment. The user interface unit 302 can select one content, and switch between diagnostic results by selecting the attribute and score of the content. For example, for "A100" with the same image ID, the score of the attribute "overall" can be compared with the score of the attribute "20 to 39 years old" and displayed.

[0040] FIG. 12 is a diagram showing an example of a screen for comparing a plurality of pieces of content in the embodiment. The user interface unit 302 allows a user to select multiple pieces of content and compare diagnostic results by selecting the attributes and scores of each piece of content.

[0041] FIG. 13 is a diagram showing an example of a screen for comparing pre-edited content and post-edited content in the embodiment. The user interface unit 302 allows the user to select the pre-correction content and the post-correction content, and to compare the diagnostic results by selecting the attributes and scores of each content.

[0042] FIG. 14 is a sequence diagram showing an example of a processing procedure of the content evaluation system 1 in the embodiment. First, the user terminal device 300 transmits the content S10 to be diagnosed to the application server device 100, and the acquisition unit 102 acquires the content S10 (step ST10). The acquisition unit 102 checks whether the uploaded content S10 satisfies a predetermined condition (step ST12). The predetermined condition is, for example, a condition for inspecting whether the content S10 is suitable for predicting an evaluation result, such as the size of the content S10. The application server device 100 transmits the content S10 to be diagnosed to the AI ​​server device 200.

[0043] The reception unit 202 of the AI ​​server device 200 receives the content S10 to be diagnosed, and the prediction unit 204 inputs the content S10 to be diagnosed to the machine learning model 204A and predicts (calculates) a biosignal based on the output of the machine learning model 204A (step ST14). The AI ​​server device 200 transmits the predicted biosignal S12 to the application server device 100. The evaluation unit 106 evaluates the content based on the biosignal S12 (step ST16). The evaluation unit 106 transmits the evaluation result S14 to the user terminal device 300. The evaluation result S14 includes feature amount information of the content S10 to be diagnosed, a desirability score and a heat map related to desirability, an interest score, a memory score, a readability score, a page ratio, a page ratio, an attention score, and a gaze estimation heat map. In addition, the desirability score, the interest score, the memory score, the readability score, and the attention score are calculated for each attribute information. The attribute information is classified, for example, into general (men and women of all ages), men of all ages, women of all ages, elderly men and women, young men and women, elderly men, young men, elderly women, and young women.

[0044] The user interface unit 302 displays the evaluation result S14, thereby allowing the user to view the evaluation of the diagnosis target content S10 (step ST18).

[0045] FIG. 15 is a sequence diagram showing another example of the processing procedure of the content evaluation system 1 in the embodiment. The user terminal device 300 transmits information S20 specifying a new comparison group to the application server device 100. The information S20 specifying the new comparison group includes IDs of multiple contents included in the new comparison group. In response to receiving the information S20 specifying the new comparison group, the application server device 100 registers the ID of the new comparison group and multiple content IDs (step ST20). The application server device 100 refers to the biosignal as the calculation result of the contents included in the new comparison group, and transmits an evaluation result S22 based on the biosignal to the user terminal device 300 (step ST22). This allows the user interface unit 302 to display the evaluation result S22 of the contents included in the new comparison group.

[0046] The user terminal device 300 transmits information S26 requesting an existing comparison group to the application server device 100. The information S26 requesting the existing comparison group includes a group ID indicating the existing comparison group. In response to receiving the information S26 requesting the existing comparison group, the application server device 100 refers to the information of the existing comparison group (step ST24), refers to the calculation result of the content included in the existing comparison group (step ST26), and transmits the comparison result S26 to the user terminal device 300. This allows the user interface unit 302 to display the comparison result S26 of the content included in the existing comparison group.

[0047] The user terminal device 300 transmits information S28 requesting addition of content to be compared or deletion of content to the application server device 100. The information S28 requesting addition of content to be compared or deletion of content includes the ID of the content to be added or deleted. In response to receiving the information S28 requesting addition of content to be compared or deletion of content, the application server device 100 refers to the information of the existing comparison group (step ST28), refers to the calculation result of the content included in the existing comparison group (step ST30), and transmits the result S30 of the content addition or deletion to the user terminal device 300. This allows the user interface unit 302 to display the comparison result of the content included in the existing comparison group while adding or deleting the content included in the existing comparison group.

[0048] The user terminal device 300 transmits information S32 requesting a comparison of the single content to the application server device 100. The information S32 requesting a comparison of the single content includes, for example, an ID of the pre-modification content and an ID of the post-modification content. In response to receiving the information S32 requesting a comparison of the single content, the application server device 100 refers to the calculation results of the pre-modification content and the calculation results of the post-modification content, and transmits an evaluation result S34 to the user terminal device 300 (step ST32). This allows the user interface unit 302 to display the comparison results of the single content.

[0049] The user terminal device 300 may transmit information S36 for posting a comment on the evaluation of the content to the application server device 100. The application server device 100, for example, registers a comment on the evaluation of each piece of content (step ST34) and transmits a comment registration result S38 to the user terminal device 300. This enables the user terminal device 300 to register a comment on the evaluation of the content, the result of comparison between contents, etc., to the application server device 100.

[0050] (Effects of the embodiment) As described above, the content evaluation system 1 according to the embodiment can realize a content evaluation system including an acquisition unit 102 that acquires content, a biosignal of a user who has viewed the content, and an evaluation system that evaluates the content based on the calculated biosignal. The content evaluation system 1 can output the evaluation of the user who has viewed the content with high accuracy.

[0051] In the content evaluation system 1, the biosignal is a signal indicating at least one change among cerebral blood flow, pupil, gaze, skin potential, heart rate, subjective evaluation, and AB test when the user views the content, and the evaluation unit 106 calculates at least one score of interest, memory, attention, likability, and readability for the content. The content evaluation system 1 can evaluate the interest, memory, attention, likability, and readability for the content with high accuracy based on the change in the biosignal when viewing.

[0052] In the content evaluation system 1, the AI ​​server device 200 inputs content to a machine learning model and calculates a biological signal based on the output of the machine learning model, and the machine learning model uses as learning data the biological signal as sensor data acquired from a sensor when a user views experimental content, and learns parameters so as to output sensor data when experimental content is input. According to the content evaluation system 1, by having the user view many experimental contents and having the machine learning model learn many biological signals, it is possible to predict biological signals for new content with high accuracy.

[0053] In the content evaluation system 1, the evaluation unit 106 creates a heat map indicating the contribution to the score for each part into which the content is divided, and therefore it is possible to present to the user the parts of the content that contributed to the score.

[0054] In the content evaluation system 1, the AI ​​server device 200 inputs content to a machine learning model corresponding to the attributes of a user and calculates a biological signal based on the output of the machine learning model, and the machine learning model learns parameters so as to output sensor data when the experimental content is input by using the biological signal as learning data obtained from a sensor when a user having the attributes is made to view experimental content. According to the content evaluation system 1, by constructing a machine learning model according to the attributes selected by the user, it is possible to predict the biological signal with high accuracy according to the attributes.

[0055] In the content evaluation system 1, the AI ​​server device 200 inputs the content to a machine learning model corresponding to the medium in which the content is used, and calculates a biological signal based on the output of the machine learning model, and the machine learning model learns parameters so as to output sensor data when the experimental content is input by using the biological signal as learning data obtained from a sensor when a user is made to view experimental content used in the medium in which the content is used. According to the content evaluation system 1, by constructing a machine learning model according to the medium in which the content is used, it is possible to predict a biological signal with high accuracy according to the medium in which the content is used.

[0056] In the content evaluation system 1, the AI ​​server device 200 inputs the content into a machine learning model corresponding to the purpose of use for which the content is to be used, calculates a biological signal based on the output of the machine learning model, and the machine learning model learns parameters so as to output sensor data when the experimental content is input, using the biological signal as learning data obtained from a sensor when a user is made to view experimental content used for the purpose of use. According to the content evaluation system 1, by constructing a machine learning model according to the purpose of use for which the content is to be used, it is possible to predict the biological signal with high accuracy according to the purpose of use.

[0057] Although each embodiment and variant example has been described, these are merely examples and are not intended to be limiting. For example, any of the embodiments or variant examples, or a part of each embodiment or a part of each variant example, may be combined with one or more other embodiments or one or more other variant examples to realize one aspect of the present invention. [Explanation of symbols]

[0058] 1...content evaluation system, 100...application server device, 102...acquisition unit, 106...evaluation unit, 110...feature extraction unit, 112...interest evaluation model calculation unit, 114...memory evaluation model calculation unit, 116...attention evaluation model calculation unit, 118...likeability evaluation model calculation unit, 120...readability evaluation model calculation unit, 122...page ratio extraction model calculation unit, 124...contribution heat map creation unit, 126...gaze estimation heat map creation unit, 200...AI server device, 202...reception unit, 204...prediction unit, 204A...machine learning model, 206...learning processing unit, 208...learning data storage unit, 210...algorithm storage unit, 300...user terminal device, 302...user interface unit, 304...application processing unit, 306...content storage unit, 400...experimental content storage unit, 402...content playback device, 404...sensor system

Claims

1. An acquisition unit that acquires content; A calculation unit that calculates a biological signal of a user who has viewed the content; an evaluation unit that evaluates the content based on the biological signal calculated by the calculation unit; A content rating system comprising:

2. the biological signal is a signal indicating at least one change among a cerebral blood flow, a pupil, a gaze, a skin potential, a heart rate, a subjective evaluation, and an AB test when the user views the content, The evaluation unit calculates at least one score of interest, memory, attention, likeability, and readability for the content. The content rating system of claim 1 .

3. The calculation unit inputs the content into a machine learning model, and calculates the biological signal based on an output of the machine learning model; The machine learning model uses a biological signal as sensor data obtained from a sensor when a user watches experimental content as learning data, and learns parameters so as to output sensor data when the experimental content is input. The content rating system of claim 1 .

4. The content evaluation system according to claim 3 , wherein the evaluation unit includes a heat map creation unit that creates a heat map indicating the contribution of each of the divided parts of the content to at least one score of interest, memory, attention, likability, and readability for the content.

5. The calculation unit inputs the content into a machine learning model corresponding to an attribute of a user, and calculates the biological signal based on an output of the machine learning model; The machine learning model uses a biological signal as sensor data obtained from a sensor when a user having the attribute is made to view experimental content as learning data, and learns parameters so as to output sensor data when the experimental content is input. The content rating system of claim 1 .

6. The calculation unit inputs the content into a machine learning model corresponding to a medium in which the content is used, and calculates the biological signal based on an output of the machine learning model; The machine learning model uses a biological signal as sensor data obtained from a sensor when the user is made to view experimental content used in the utilization medium as learning data, and learns parameters so as to output sensor data when the experimental content is input. The content rating system of claim 1 .

7. The calculation unit inputs the content into a machine learning model corresponding to a purpose of use of the content, and calculates the biological signal based on an output of the machine learning model; The machine learning model uses a biological signal as sensor data obtained from a sensor when the user is made to view experimental content used for the intended use as learning data, and learns parameters so as to output sensor data when the experimental content is input. The content rating system of claim 1 .

8. An information processing device acquires content; A step in which the information processing device calculates a biological signal of a user who has viewed the content; The information processing device evaluates the content based on the calculated biosignal; A content evaluation method, including:

9. The computer of the information processing device obtaining content; calculating a biological signal of a user who has viewed the content; evaluating the content based on a computed biosignal; A program to execute.

10. A reception unit that receives content; A prediction unit that inputs the content into a machine learning model and predicts a biological signal based on an output of the machine learning model, The machine learning model uses a biological signal as sensor data obtained from a sensor when a user watches experimental content as learning data, and learns parameters so as to output sensor data when the experimental content is input. Server device.

Images