Emotion estimation device and emotion estimation method

Abstract

The emotion prediction device includes a biological information acquisition unit, an emotion information acquisition unit, an association construction unit, an emotion prediction acquisition unit, and a prediction unit. The biological information acquisition unit acquires and parses first biological information representing the physical state of each of the multiple subjects within a preparation space. The emotion information acquisition unit acquires emotion information representing the emotions of each of the multiple subjects. The association construction unit constructs association information that links the first biological information and the corresponding emotion information. The emotion prediction acquisition unit acquires second biological information representing the physical state of any subject within the prediction space. The prediction unit extracts emotion information from the association information that is associated with the first biological information equal to the second biological information, and predicts the emotion of the arbitrary subject based on the extracted emotion information.

Description

Technical Field

[0001] The present disclosure relates to an emotion estimation device and an emotion estimation method for estimating a person's sense of comfort. Background Art

[0002] Devices for estimating human emotions are known. For example, Patent Document 1 describes an emotion determination device that estimates brain activity of a subject based on acquired facial skin temperature data or facial blood circulation volume data, and determines the subject's emotions based on the estimated brain activity.

[0003] Specifically, the emotion determination device has an emotion storage unit that associates and stores the estimated result of brain activity with emotion data representing the emotion of the person. The emotion determination device determines the emotion of the subject by extracting the emotion data associated with the brain activity estimated based on the facial skin temperature data or the facial blood circulation volume data in the emotion storage unit. In addition, for example, based on the data of the amount of brain activity collected when the person is placed in any state of "joy, anger, sorrow, or happiness" and a brain function activation task is provided to the person for a certain period of time, the estimated result of brain activity stored in the emotion storage unit is associated with the emotion data.

[0004] Prior art literature

[0005] Patent Document 1: Japanese Patent Application Publication No. 2017-74214 Summary of the invention

[0006] Here, the above-mentioned emotion determination device collects data on brain activity after putting a person in any state of "joy, anger, sorrow, or happiness" by some method, and associates the emotion data with the estimated results of brain activity. That is, after a person's emotions are changed by some external stimulation, a database is constructed that associates biological information such as emotion information representing the changed emotions and estimated results of brain activity after the emotion change. However, in the case of obtaining a database used for emotion estimation through experiments such as providing specific stimulation to people or causing specific environmental factors to change, there is a possibility that emotions cannot be estimated with high accuracy due to differences between the experimental environment and the daily environment, or the influence of individual differences.

[0007] The present disclosure has been made to solve the above-mentioned problems, and an object of the present disclosure is to provide an emotion estimation device and an emotion estimation method capable of improving the estimation accuracy of the emotion of an estimation target person.

[0008] An emotion estimation device involved in one aspect of the present disclosure is an emotion estimation device for estimating emotions of multiple subjects existing in an estimation space, comprising: a biological information acquisition unit for acquiring and analyzing first biological information indicating the physical state of each of the multiple subjects existing in the preparation space; an emotional information acquisition unit for acquiring emotional information indicating the emotions of each of the multiple subjects existing in the preparation space; an association construction unit for constructing association information that associates the first biological information and the emotional information corresponding to the first biological information with each other; an estimated biological information acquisition unit for acquiring and analyzing second biological information indicating the physical state of any subject existing in the estimation space among the multiple subjects; and an estimation unit for extracting the emotional information associated with the first biological information that is equal to the second biological information in the association information, and estimating the emotion of the arbitrary subject existing in the estimation space based on the extracted emotional information.

[0009] An emotion inference method according to one aspect of the present disclosure is an emotion inference method performed by an emotion inference device for inferring emotions of a plurality of subjects existing in an inference space, comprising: a biometric information acquisition step of acquiring and analyzing first biometric information indicating the state of a body of each of the plurality of subjects existing in the preparation space; an emotion information acquisition step of acquiring emotion information indicating the emotion of each of the plurality of subjects existing in the preparation space; an association construction step of constructing association information associating the first biometric information and the emotion information corresponding to the first biometric information with each other; an inferred biometric information acquisition step of acquiring and analyzing second biometric information indicating the state of a body of any subject existing in the inference space among the plurality of subjects; and an inference step of extracting the emotion information associated with the first biometric information equal to the second biometric information in the association information, and inferring the emotion of the arbitrary subject existing in the inference space based on the extracted emotion information.

[0010] The present disclosure can provide an emotion estimation device and an emotion estimation method that can improve the estimation accuracy of the emotion of an estimation target person. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 This is a diagram illustrating functional blocks of the emotion estimation device according to the first embodiment.

[0012] Figure 2 This is a schematic diagram illustrating the hardware configuration of the emotion estimation device according to the first embodiment.

[0013] Figure 3This is a flowchart illustrating the emotion estimation process of the emotion estimation device according to the first embodiment.

[0014] Figure 4 This is a flowchart illustrating the emotion estimation process involved in the second embodiment.

[0015] Figure 5 This is a flowchart illustrating the emotion estimation process involved in the third embodiment. DETAILED DESCRIPTION

[0016] Hereinafter, an emotion estimation device and an emotion estimation method according to an embodiment will be described in detail with reference to the drawings.

[0017] Implementation method 1.

[0018] The emotion estimation device 100 according to the first embodiment is a device that estimates the emotion of a subject in a specific indoor or indoor space such as a room in an office building based on the subject's biological information described later. Hereinafter, the specific indoor or indoor space may be referred to as an estimation space. Figure 1 This is a diagram illustrating functional blocks of the emotion estimation device according to Embodiment 1. The emotion estimation device 100 includes a storage unit 1 , a prepared biometric information acquisition unit 2 , a prepared emotion information acquisition unit 3 , a correlation unit 4 , an estimated biometric information acquisition unit 5 , an estimation unit 6 , and an output unit 7 .

[0019] The storage unit 1 stores related information used by the emotion estimation device 100 to estimate the emotions of the subject. The emotion estimation device 100 refers to the related information when estimating the emotions of the subject.

[0020] The preparation biological information acquisition unit 2 includes one or more sensors. Hereinafter, the sensor may be described as a preparation sensor. The preparation biological information acquisition unit 2 uses one or more preparation sensors to acquire the first biological information of the sample from each of the plurality of subjects present in the preparation space. The preparation space refers to a space such as a laboratory for acquiring the first biological information from the plurality of subjects. The preparation space may be the same space as the above-mentioned estimated space, a space having the same indoor purpose as the estimated space, or a space similar to the estimated space. In addition, the space similar to the estimated space may be a space whose temperature difference with the estimated space is less than a predetermined temperature difference threshold, or a space whose humidity difference with the estimated space is less than a predetermined humidity difference threshold. Alternatively, the similar space may be a space whose illumination difference with the estimated space is less than a predetermined illumination difference threshold, or a space whose area difference with the estimated space is less than a predetermined area difference threshold. In addition, the above-mentioned similar space is only an example, and the space similar to the estimated space is not limited to the above-mentioned example. In addition, the above-mentioned preparation space is only an example, and the preparation space is not limited to the above-mentioned example. When the preparation space is the same space as the estimated space, the space where the indoor purpose is the same as the indoor purpose of the estimated space, or the space is similar to the estimated space, the emotion estimation device 100 can obtain the first biometric information without placing a special burden on the above-mentioned multiple subjects. On the other hand, when the preparation space is not the same space as the estimated space, the space where the indoor purpose is the same as the indoor purpose of the estimated space, or the space is similar to the estimated space, the first biometric information can be easily obtained using the emotion estimation device 100. The first biometric information is information that becomes an indicator of the state of the human body. The first biometric information is represented by a numerical value and becomes various values ​​according to the state of the human body. The first biometric information is, for example, information indicating at least one of the body surface temperature, heartbeat, pulse, skin galvanic response, the time variation of the body surface temperature, the difference in the surface temperature of each of the multiple parts of the body, and the index value obtained by analyzing the frequency of the heartbeat. The preparation biometric information acquisition unit 2 in the first embodiment, for example, uses the above-mentioned one or more preparation sensors to obtain the first biometric information indicating at least any one of the body surface temperature, heartbeat, pulse, and skin galvanic response.

[0021] The biological information acquisition unit 2 may also acquire the first biological information from one part of the human body by any one detection method. Alternatively, the biological information acquisition unit 2 may also acquire one kind of first biological information or each first biological information of two or more first biological information from one part of the human body by any two or more detection methods. Alternatively, the biological information acquisition unit 2 may also acquire one kind of first biological information or each first biological information of two or more first biological information from multiple parts of the human body by any two or more detection methods.

[0022] The preparation biological information acquisition unit 2 in the first embodiment analyzes the first biological information acquired by the above-mentioned one or more preparation sensors. Specifically, the preparation biological information acquisition unit 2 uses the first biological information acquired by the above-mentioned one or more preparation sensors to generate the first biological information indicating at least one of the temporal variation of the body surface temperature, the difference in the surface temperature of each of the plurality of parts in the body, and the index value obtained by the frequency analysis of the heartbeat. For example, when the first biological information acquired by the one or more preparation sensors is the body surface temperature, the preparation biological information acquisition unit 2 may calculate the difference in the surface temperature of each of the plurality of parts in the body. Alternatively, when the first biological information acquired by the one or more preparation sensors is data related to the heartbeat, such as electrocardiogram data, the preparation biological information acquisition unit 2 may perform frequency analysis to acquire the index value indicating the state of the heartbeat.

[0023] The preparation emotion information acquisition unit 3 acquires emotion information representing the emotions of each of the plurality of subjects present in the preparation space. Here, the emotion information is information representing emotions such as comfort, excitement, or calmness using numerical values. The preparation emotion information acquisition unit 3 includes an input device such as a keyboard, a touch panel, a button, or a mouse for each of the plurality of subjects to input emotion information to the emotion estimation device 100.

[0024] The preparation emotion information acquisition unit 3 in Embodiment 1 acquires emotion information from each of the plurality of subjects existing in the preparation space according to the experience extraction method. That is, the preparation emotion information acquisition unit 3 acquires emotion information representing the emotions of each of the plurality of subjects at each of the plurality of time points within a predetermined period from each of the plurality of subjects existing in the preparation space.

[0025] The emotion information acquisition unit 3 is prepared to acquire emotion information at each of a plurality of time points in the above-mentioned period of, for example, more than 2 days, based on an experience extraction method. The emotion information acquisition unit 3 is prepared to acquire emotion information at one or more predetermined moments or random moments on each date within the period. The emotion information acquisition unit 3 is preferably prepared to acquire emotion information in a period of more than 5 days. The reason for this is that there are many situations where different events occur on each working day or each day in a week, and the emotion information acquisition unit 3 is prepared to acquire emotion information indicating the physical and mental state caused by the event without omission.

[0026] The emotion information obtained at each of the multiple time points by the experience extraction method is information in which the recall bias of each subject is reduced. That is, the emotion information obtained at each time point is information that accurately represents the emotion of each subject affected by the event that occurred at each time point. In addition, the set of emotion information obtained at each of the multiple time points by the experience extraction method is not special information that only represents a specific emotion caused by a specific event in daily life, but is information that represents various emotions of each subject caused by various events.

[0027] The association construction unit 4 constructs association information that associates the first biometric information analyzed by the preparation biometric information acquisition unit 2 and the emotion information corresponding to the first biometric information. Here, the emotion information corresponding to the first biometric information is information that expresses the above-mentioned emotions such as comfort, excitement or calmness, obtained by using the emotion information acquired from each subject by the preparation emotion information acquisition unit 3 at the same time point as the time point when the preparation biometric information acquisition unit 2 acquires the first biometric information from each subject using one or more preparation sensors. In addition, the time point that is set to be the same as a certain time point refers to a time point within a predetermined time before and after the certain time point as the starting point. The predetermined time refers to a time when the emotional state and physical state of a person obtained in advance through experiments, etc. do not change.

[0028] The emotion information corresponding to the first biometric information may be the emotion information acquired at the same time point as the first biometric information, may be the information obtained by regression analysis, or may be the information obtained by averaging the value of the emotion information for each first biometric information. Hereinafter, an example is given in which the first biometric information is the amount of change in body surface temperature over time. For example, in the case where the first biometric information indicating 0.5[°C] is acquired from a plurality of subjects, the value of the emotion information acquired from the plurality of subjects at the same time point as the time point when the 0.5[°C] is acquired, or the average value of the value of the emotion information becomes the example of "emotion information corresponding to the first biometric information indicating 0.5[°C]". In addition, in the case where the relationship between the value of the first biometric information and the value of the emotion information is formulated by regression analysis using the first biometric information and the values ​​of the emotion information acquired from the plurality of subjects, the value of the emotion information obtained when the 0.5[°C] is substituted into the formula becomes the example of "emotion information corresponding to the first biometric information indicating 0.5[°C]".

[0029] The emotion information corresponding to the first biometric information is information obtained by generalizing the emotions of the plurality of subjects at the same time point as the time point when the first biometric information was obtained, with respect to the first biometric information indicating the physical state of the plurality of subjects. For example, when the emotion information of a numerical value indicating a high level of comfort is obtained from the plurality of subjects at all the time points when the first biometric information indicating the above-mentioned 0.5[℃] is obtained from the plurality of subjects, the "emotion information corresponding to the first biometric information indicating the 0.5[℃]" becomes a numerical value indicating a high level of comfort. It is considered that the high level of comfort generally indicates the emotions of the plurality of subjects at the time point when the first biometric information indicating the 0.5[℃] is obtained.

[0030] The associated information may also be a formula that associates the first biometric information with the emotion information corresponding to the first biometric information, for example, using the above-mentioned regression analysis. Alternatively, the associated information may also be information that represents the value of the first biometric information and the value of the emotion information corresponding to the first biometric information, for example, in a table format. In addition, the associated information is not limited to the above-mentioned examples as long as it is information that associates the first biometric information with the emotion information corresponding to the first biometric information.

[0031] The estimated biometric information acquisition unit 5 includes one or more sensors. Hereinafter, the sensor is sometimes described as an estimation sensor. The one or more estimation sensors may also be the same sensors as the one or more preparation sensors mentioned above. The estimated biometric information acquisition unit 5 uses the one or more estimation sensors to acquire the second biometric information from any of the multiple subjects present in the estimation space. In addition, as described above, the estimation space refers to a space used by the emotion estimation device 100 to estimate the emotions of each of the multiple subjects. The estimation space may be the same space as the preparation space mentioned above, or a different space. The second biometric information is information that is an indicator of the state of the human body. The second biometric information is represented by a numerical value and becomes various values ​​according to the state of the human body. The second biometric information is, for example, information indicating at least one of body surface temperature, heartbeat, pulse, skin galvanic response, the amount of change in body surface temperature over time, the difference in surface temperature of each of multiple parts of the body, and an index value obtained by analyzing the frequency of the heartbeat. The estimated biological information acquisition unit 5 in the first embodiment acquires the second biological information indicating at least one of body surface temperature, heartbeat, pulse, and galvanic skin response, for example, using the one or more estimation sensors.

[0032] The estimated biological information acquisition unit 5 may also acquire biological information from one part of the human body by any one detection method. Alternatively, the estimated biological information acquisition unit 5 may also acquire one kind of biological information or various biological information of two or more biological information from one part of the human body by any two or more detection methods. Alternatively, the estimated biological information acquisition unit 5 may also acquire one kind of biological information or various biological information of two or more biological information from multiple parts of the human body by any two or more detection methods.

[0033] The estimated biometric information acquisition unit 5 in the first embodiment analyzes the second biometric information acquired by the one or more estimation sensors. Specifically, the estimated biometric information acquisition unit 5 generates the second biometric information indicating at least one of the temporal variation of the body surface temperature, the difference in the surface temperature of each of the plurality of parts of the body, and the index value obtained by analyzing the frequency of the heartbeat, using the second biometric information acquired by the one or more estimation sensors. In addition, the second biometric information generated by the estimated biometric information acquisition unit 5 is the same type of information as the first biometric information included in the associated information. In the case where the first biometric information included in the associated information is, for example, the temporal variation of the body surface temperature, the second biometric information generated by the estimated biometric information acquisition unit 5 becomes the temporal variation of the body surface temperature.

[0034] The estimation unit 6 extracts the emotion information associated with the first biometric information equal to the second biometric information analyzed by the estimated biometric information acquisition unit 5, using the associated information. Here, when the associated information is information that represents the value of the first biometric information and the value of the emotion information corresponding to the first biometric information, for example, in a table format, the first biometric information equal to the second biometric information is assumed to be the biometric information with the smallest difference with the second biometric information. The estimation unit 6 estimates the emotion represented by the extracted emotion information as the emotion of the subject who is the acquisition source of the second biometric information.

[0035] The output unit 7 communicates with external devices. As the external device, for example, a personal computer, a workstation, a tablet terminal, or a smart phone and other terminals can be listed. In addition, as the external device, for example, a display device such as a liquid crystal display, a CRT (Cathode Ray Tube), or an organic EL (Electro Luminescence) display can be listed. In addition, as the external device, for example, an environment adjustment device such as an air conditioner or an air purifier that adjusts the environment of the estimated space can be listed. In addition, as the external device, a sound output device including a speaker can be listed.

[0036] When the external device is a terminal, the inference unit 6 controls the output unit 7 in such a manner as to send the extracted emotion information to the terminal. When the external device is a display device, the inference unit 6 controls the output unit 7 in such a manner as to send the emotion information and display instruction information indicating the display of the emotion information to the display device. When the external device is an environment adjustment device, the inference unit 6 controls the output unit 7 in such a manner as to send the emotion information and an instruction for causing the environment adjustment device to perform an action corresponding to the emotion information to the environment adjustment device. In addition, the action of the environment adjustment device corresponding to the emotion information refers to an action that improves the emotional state of the subject in the inference space. When the external device is a sound output device, the inference unit 6 controls the output unit 7 in such a manner as to send the emotion information and an instruction for outputting the emotion information by sound to the sound output device.

[0037] Hereinafter, the hardware configuration of the emotion estimation device 100 according to the first embodiment will be described. Figure 21 is a schematic diagram illustrating the hardware structure of the emotion estimation device involved in Implementation Example 1. The emotion estimation device 100 may include, for example, a processor 11, a memory 12, a storage device 13, a first input interface circuit 14, a sensor 15, a second input interface circuit 16, an input device 17, and an output interface circuit 18. In addition, the processor 11, the memory 12, the storage device 13, the first input interface circuit 14, the second input interface circuit 16, and the output interface circuit 18 are connected via a bus 19.

[0038] As the processor 11, for example, a CPU (Central Processing Unit) or an MPU (Micro Processing Unit) can be listed. As the memory 12, for example, a ROM (Read Only Memory) or a RAM (Random Access Memory) can be listed. As the storage device 13, for example, an HDD (Hard Disk Drive) or an SSD (Solid State Drive) can be listed. As the sensor 15, for example, an infrared sensor or a wearable sensor can be listed. As the input device 17, for example, a keyboard or a touch panel can be listed. The first input interface circuit 14 is an input interface circuit that bridges the processor 11 and the sensor 15. The second input interface circuit 16 is an input interface circuit that bridges the processor 11 and the input device 17.

[0039] The functions of the storage unit 1 can be realized by the storage device 13. The functions of the association construction unit 4 and the inference unit 6 can be realized by the processor 11 reading and executing various programs such as an emotion inference program for inferring emotions stored in the memory 12 or the storage device 13. The acquisition function of the first biometric information of the biometric information acquisition unit 2 and the acquisition function of the second biometric information of the biometric information acquisition unit 5 can be realized by the first input interface circuit 14 and the sensor 15. The analysis function of the first biometric information of the biometric information acquisition unit 2 and the analysis function of the second biometric information of the biometric information acquisition unit 5 can be realized by the processor 11 reading and executing various programs such as the above-mentioned emotion inference program stored in the memory 12 or the storage device 13. The function of preparing the emotion information acquisition unit 3 to acquire emotion information at each of a plurality of time points can be realized by the processor 11, the second input interface circuit 16 and the input device 17. Furthermore, the memory 12 or the storage device 13 can realize a function in which the association building unit 4 or the prepared emotion information acquisition unit 3 maintains the emotion information acquired from each of the plurality of subjects by the prepared emotion information acquisition unit 3 at each of the plurality of time points until the association building unit 4 generates the association information. The memory 12 or the storage device 13 can realize a function in which the association building unit 4 or the prepared biometric information acquisition unit 2 maintains the first biometric information acquired and analyzed from each of the plurality of subjects by the prepared biometric information acquisition unit 2 at each of the plurality of time points until the association building unit 4 generates the association information.

[0040] The function of the output unit 7 can be realized by the output interface circuit 18. In addition, when the external device is a terminal, the output interface circuit 18 is a communication interface circuit. On the other hand, when the external device is any device among a display device, a sound output device, and an environment adjustment device, the output interface circuit 18 is an output interface circuit that bridges the arbitrary device and the processor 11.

[0041] The emotion estimation device 100 may also include a plurality of the first input interface circuits 14 and the sensors 15, respectively. Thus, when a plurality of subjects are located at different positions in the preparation space and the estimation space, respectively, the emotion estimation device 100 can simultaneously obtain the first biological information and the second biological information from each of the plurality of subjects. The emotion estimation device 100 may also include a plurality of the second input interface circuits 16 and the input devices 17, respectively. Thus, when a plurality of subjects are located at different positions in the preparation space, the emotion estimation device 100 can simultaneously obtain the emotion information from each of the plurality of subjects. All or part of the functions of the emotion estimation device 100 may also be implemented by dedicated hardware.

[0042] Next, refer to Figure 3 , describing the emotion estimation processing of the emotion estimation device 100 according to the first embodiment. Figure 3 This is a flowchart illustrating the emotion estimation processing of the emotion estimation device involved in Implementation Example 1. In step S1, the emotion information acquisition unit 3 determines whether the current time point is the timing for acquiring emotion information. Hereinafter, this timing is recorded as the acquisition timing. When the current time point is not the acquisition timing (step S1: "No"), the emotion information acquisition unit 3 returns the emotion estimation processing to step S1.

[0043] When the current time point is the acquisition timing (step S1: "Yes"), in step S2, the emotion information acquisition unit 3 is prepared to acquire emotion information from each of the multiple subjects. In step S3, the biometric information acquisition unit 2 is prepared to analyze the first biometric information acquired from each subject at the same time point as the acquisition time point of the emotion information. In addition, the biometric information acquisition unit 2 may acquire the first biometric information from each subject at the same time point as the acquisition time point of the emotion information, or may continuously acquire the first biometric information from each subject. In the case where the biometric information acquisition unit 2 continuously acquires the first biometric information from each subject, it may also analyze the continuously acquired first biometric information.

[0044] In step S4, the preparation emotion information acquisition unit 3 determines whether the above period has ended. If the period has not ended (step S4: "No"), the preparation emotion information acquisition unit 3 returns the emotion estimation process to step S1. If the above period has ended (step S4: "Yes"), in step S5, the association construction unit 4 uses the emotion information acquired during the period and the first biometric information to construct association information that associates the first biometric information with the emotion information corresponding to the first biometric information. In step S6, the association construction unit 4 stores the constructed association information in the storage unit 1.

[0045] In step S7, the estimated biometric information acquisition unit 5 acquires the second biometric information of any of the above-mentioned subjects among the multiple subjects. In step S8, the estimated biometric information acquisition unit 5 analyzes the acquired second biometric information. In step S9, the estimation unit 6 refers to the associated information and extracts the emotion information associated with the first biometric information that is equal to the analyzed second biometric information. The estimation unit 6 estimates the emotion represented by the extracted emotion information as the emotion of the subject. In step S10, the estimation unit 6 instructs the output unit 7 to output the extracted emotion information to an external device. The output unit 7 outputs the emotion information to an external device in accordance with the instruction from the estimation unit 6. After the processing of step S10, the emotion estimation processing is terminated. In addition, when the emotion estimation processing is executed again, the emotion estimation device 100 starts from the processing of step S7.

[0046] The effect of the emotion estimation device 100 involved in the first embodiment is described below. The emotion estimation device 100 involved in the first embodiment includes a prepared biometric information acquisition unit 2, a prepared emotion information acquisition unit 3, a correlation construction unit 4, an estimated biometric information acquisition unit 5, and an estimation unit 6. The prepared biometric information acquisition unit 2 acquires and analyzes the first biometric information indicating the state of the body of each of the plurality of subjects existing in the prepared space. The prepared emotion information acquisition unit 3 acquires the emotion information indicating the emotion of each of the plurality of subjects existing in the prepared space. The correlation construction unit 4 constructs the association information associating the first biometric information and the emotion information corresponding to the first biometric information. The estimated biometric information acquisition unit 5 acquires and analyzes the second biometric information indicating the state of the body of any subject among the plurality of subjects existing in the estimation space. The estimation unit 6 extracts the emotion information associated with the first biometric information equal to the second biometric information from the association information constructed by the correlation construction unit 4. Then, the estimation unit 6 estimates the emotion of the arbitrary subject existing in the estimation space based on the extracted emotion information. As a result, the accuracy of emotion estimation is improved.

[0047] The association construction unit 4 in the first embodiment uses the emotion information obtained by the emotion information acquisition unit 3 from each of the multiple subjects and the first biometric information obtained by the biometric information acquisition unit 2 from each of the multiple subjects at the same time point as the time point when the emotion information was obtained, to construct the association information. In addition, the same time point as the time point when the emotion information was obtained refers to a time point within a predetermined time before and after the time point when the emotion information acquisition unit 3 obtained the emotion information from each of the multiple subjects. As a result, the emotion information and the first biometric information obtained immediately before or after the time point when the emotion information was obtained are reflected in the association information, and the accuracy of the association information is improved. Therefore, the accuracy of the emotion estimation obtained by the emotion estimation device 100 using the association information is improved.

[0048] The preparation emotion information acquisition unit 3 in the first embodiment acquires emotion information representing the emotion of each of the plurality of subjects at each of the plurality of time points within a predetermined period from each of the plurality of subjects present in the preparation space. The association construction unit 4 constructs the association information using the first biometric information acquired from each of the plurality of subjects by the preparation biometric information acquisition unit 2 at each of the plurality of time points and the emotion information acquired from each of the plurality of subjects by the preparation emotion information acquisition unit 3 at each of the plurality of time points. As a result, the recall bias of the emotion information acquired by the preparation emotion information acquisition unit 3 is reduced. In addition, the correlation between the first biometric information and the emotion information corresponding to the first biometric information in the association information does not depend on a specific event such as a climate change, a change in physical condition, or a change in mood. Therefore, with respect to the association information, the influence caused by the event is reduced, and the accuracy is improved. Therefore, the accuracy of the emotion estimation by the estimation unit 6 is improved.

[0049] The above-mentioned period in the first embodiment is a period of more than 5 days. The prepared emotion information acquisition unit 3 acquires emotion information from each of the plurality of subjects at one or more predetermined moments or at one or more random moments on each date. Thus, the associated information does not depend on the weather, physical condition, or mood of each date on a 5-day period of a working day, each date on a weekday, or each season. Therefore, the influence of various events on the associated information is reduced, and the accuracy is improved. Therefore, the accuracy of the emotion estimation by the estimation unit 6 is improved.

[0050] The emotion estimation device 100 involved in embodiment 1 also has an output unit 7 for communicating with an external device. The estimation unit 6 controls the output unit 7 in a manner that sends the extracted emotion information to the external device. Thus, in the case where the external device is a terminal, a display device, or a sound output device, the user of the terminal, the display device, or the sound output device can know the emotions of the subject. In addition, in the case where the external device is the above-mentioned environment adjustment device, the environment adjustment device can improve the comfort level of the subject, or improve the mental productivity of the subject related to the comfort level, by performing actions corresponding to the emotions of the subject.

[0051] Implementation method 2.

[0052] The emotion estimation device 100 involved in the second embodiment constructs association information that associates the first biological information and the emotion information corresponding to the first biological information for each subject. And, by using the association information, the emotion estimation device 100 estimates the emotions of each subject with higher accuracy. The second embodiment is described below. In addition, among the components in the second embodiment, the same symbols as those in the first embodiment are attached to the components that are the same as the components in the first embodiment. In addition, in the content of the second embodiment, the description of the same content as that of the first embodiment is omitted unless there are special circumstances.

[0053] The functional blocks included in the emotion estimation device 100 according to the second embodiment are Figure 1 The functional blocks in the emotion estimation device 100 according to the first embodiment are the same.

[0054] The prepared biometric information acquisition unit 2 in the second embodiment includes a sensor such as a wearable sensor assigned to each of the plurality of subjects. Sensor identification information for identifying the sensor is assigned to the sensor assigned to each subject. Alternatively, the prepared biometric information acquisition unit 2 in the second embodiment includes a camera including a sensor for recognizing the face of each of the plurality of subjects. Thus, the prepared biometric information acquisition unit 2 recognizes each subject. In addition, the prepared biometric information acquisition unit 2 is not limited to the above example as long as it can recognize each subject.

[0055] The prepared emotion information acquisition unit 3 in the second embodiment includes, for example, a terminal assigned to each of the multiple subjects. An ID (Identifier) ​​for identifying the terminal is assigned to the terminal. Alternatively, the prepared emotion information acquisition unit 3 in the second embodiment acquires emotion information and identification information for identifying each subject from each subject. It is assumed that in the emotion estimation device 100, the ID of the terminal of each subject or the identification information of each subject, etc., are associated with the sensor identification information of the sensor assigned to each subject or the facial data of each subject. Hereinafter, the information for identifying each subject, such as the ID of the terminal of each subject, the above-mentioned identification information of each subject, the sensor identification information of the sensor assigned to each subject, and the facial data of each subject, is collectively referred to as identification information.

[0056] The association construction unit 4 in the second embodiment generates association information that associates the first biometric information indicating the body of the subject person with the emotion information indicating the emotion of the subject person and corresponding to the first biometric information for each identification information of the subject person. That is, the association construction unit 4 generates association information that associates the identification information of each subject person, the first biometric information indicating the state of the body of each subject person, and the emotion information corresponding to the first biometric information indicating the state of the body of each subject person.

[0057] Here, the emotion information corresponding to the first biometric information indicating the state of the body of each subject may be the emotion information obtained from each subject at the same time point as the first biometric information. Alternatively, it may be information obtained by the above-mentioned regression analysis or the above-mentioned average calculation process using the emotion information obtained from each subject at the same time point as the first biometric information. Hereinafter, an example is given in which the first biometric information is the temporal variation of the body surface temperature. It is assumed that the emotion information obtained from a certain subject at the same time point as the time point when the first biometric information indicating 0.5 [°C] is obtained indicates a high level of comfort. In this case, the numerical value indicating the high level of comfort or the result obtained by averaging the numerical value becomes an example of "emotion information corresponding to the temporal variation of the body surface temperature of 0.5 [°C] of the subject of the certain type", that is, "emotion information corresponding to the first biometric information indicating the state of the body of the certain subject". In addition, when the relationship between the first biometric information and the emotion information obtained from a certain subject is formulated by regression analysis or the like using the respective values ​​of the first biometric information and the emotion information obtained from the certain subject, the value of the emotion information obtained by substituting 0.5 [°C] into the formula becomes an example of "emotion information corresponding to the first biometric information representing the physical state of the certain subject."

[0058] The emotion information corresponding to the first biometric information indicating the state of the body of each subject is information that generalizes the emotion of each subject. For example, when the emotion information indicating a numerical value of a high comfort level is obtained from the subject at all the time points when the first biometric information indicating the above 0.5 [°C] is obtained, the first biometric information indicating the state of the body of the subject, that is, the emotion information corresponding to the temporal change in the body surface temperature of 0.5 [°C], indicates the high comfort level of the subject. It is considered that the high comfort level generally indicates the emotion of the subject at the time point when the first biometric information indicating the 0.5 [°C] is obtained.

[0059] The estimation unit 6 extracts the emotion information associated with the identification information of each subject and the first biometric information equal to the second biometric information indicating the physical state of each subject acquired by the estimated biometric information acquisition unit 5 from the associated information. Then, the estimation unit 6 estimates the emotion indicated by the extracted emotion information as the emotion of each subject.

[0060] Figure 4 This is a flowchart of the emotion estimation processing involved in the exemplary embodiment 2. The processing in step S21 is the same as the processing in the above-mentioned step S1, so the description is omitted. In addition, in step S21, when the current time point is not the acquisition timing (step S21: "No"), the prepared emotion information acquisition unit 3 returns the emotion estimation processing to step S21. In step S21, when the current time point is the acquisition timing (step S21: "Yes"), in step S22, the prepared emotion information acquisition unit 3 acquires identification information and emotion information from multiple subjects. In step S23, the prepared biometric information acquisition unit 2 analyzes the first biometric information acquired from each subject at the same time point as the time point when the emotion information was acquired from each subject. In addition, the prepared biometric information acquisition unit 2 acquires the first biometric information from each subject, and acquires the above-mentioned sensor identification information or the above-mentioned facial data and other above-mentioned identification information for identifying each subject. The preparation biometric information acquisition unit 2 may acquire the first biometric information from each subject at the same time as the time of acquiring the emotion information, or may acquire the first biometric information from each subject continuously. When the preparation biometric information acquisition unit 2 acquires the first biometric information from each subject continuously, it may analyze the first biometric information acquired continuously.

[0061] In step S24, the preparation emotion information acquisition unit 3 determines whether the above period has ended. In the case where the period has not ended (step S24: "No"), the preparation emotion information acquisition unit 3 returns the emotion estimation process to step S21. In the case where the above period has ended (step S24: "Yes"), in step S25, the association construction unit 4 constructs association information that associates the identification information of each subject, the first biometric information indicating the state of the body of each subject, and the emotion information indicating the emotion of each subject and corresponding to the first biometric information. The processing in step S26 is the same as the processing in the above step S6, so the description is omitted.

[0062] In step S27, the estimated biometric information acquisition unit 5 acquires the second biometric information from the above-mentioned arbitrary subject among the multiple subjects. At this time, the estimated biometric information acquisition unit 5 acquires the second biometric information, and acquires the above-mentioned sensor identification information or the above-mentioned identification information such as the above-mentioned facial data for identifying the arbitrary subject. The processing of step S28 is the same as the processing of step S8, so the description is omitted. In step S29, the estimation unit 6 refers to the associated information and extracts the emotion information associated with the identification information of the arbitrary subject and the first biometric information equal to the analyzed second biometric information. The estimation unit 6 estimates the emotion represented by the extracted emotion information as the emotion of the subject. The processing in step S30 is the same as the processing in step S10, so the description is omitted. After the processing of step S30, the emotion estimation processing ends. In addition, when the emotion estimation processing is executed again, the emotion estimation device 100 starts from the processing of step S27.

[0063] The following describes the effects of the emotion estimation device 100 involved in the second embodiment. The prepared biometric information acquisition unit 2 and the prepared emotion information acquisition unit 3 in the second embodiment acquire identification information for identifying each of the plurality of subjects. The association construction unit 4 constructs association information that associates the identification information of each subject, the first biometric information indicating the state of the body of each subject, and the emotion information indicating the emotion of each subject and corresponding to the first biometric information indicating the state of the body of each subject. The estimation unit 6 extracts the emotion information associated with the identification information of any subject and the first biometric information equivalent to the second biometric information indicating the state of the body of the arbitrary subject in the association information, and estimates the emotion of the arbitrary subject based on the extracted emotion information. Thus, the association information shows the correlation between the state of the body and the emotion for each subject. Therefore, the estimation unit 6 can estimate the emotion for each subject by using the association information. Therefore, even if there are differences in characteristics such as fear of heat or fear of cold for each subject, the emotion estimation device 100 can estimate the emotion according to the subject with high accuracy.

[0064] Implementation method 3.

[0065] The emotion estimation device 100 involved in the third embodiment obtains emotion information according to the evaluation grid method in the construction of the estimation information, and classifies the object person by performing cluster analysis. The third embodiment is described below. In addition, among the components in the third embodiment, the components that are the same as the components in the first and second embodiments are given the same symbols as those in the first and second embodiments. In addition, in the content of the third embodiment, the description of the same contents as those in the first and second embodiments is omitted unless there are special circumstances.

[0066] The functional blocks included in the emotion estimation device 100 according to the third embodiment are Figure 1 Functional blocks in the emotion estimation device 100 according to the first and second embodiments described above are the same.

[0067] The preparation emotion information acquisition unit 3 in the third embodiment acquires emotion information from each of the plurality of subjects according to the evaluation grid method. In this case, the preparation emotion information acquisition unit 3 acquires the comfort level, which is an index of the comfort level, as emotion information according to the scoring method. In addition, the preparation emotion information acquisition unit 3 acquires factor information indicating the factors of the comfort level together with the comfort level according to the free descriptive method.

[0068] The association construction unit 4 uses the comfort and factor information obtained by the preparation emotion information acquisition unit 3 to perform cluster analysis. Specifically, the association construction unit 4 uses the factor information from each subject to classify each subject into any type among a plurality of types. For example, the association construction unit 4 classifies a subject whose main cause of comfort is the warmth factor as a "warm type" or the like. In addition, the association construction unit 4 classifies a subject whose main cause of comfort is an internal factor such as the state of the body and mind as an "inner type" or the like. Moreover, the association construction unit 4 classifies a subject whose comfort is affected by various factors as a "balanced type" or the like. Here, what factors affect the comfort of each subject is determined by the factor information obtained from the subject. For example, if the factor information of a predetermined ratio or more in the factor information from a certain subject is content related to the temperature, the association construction unit 4 classifies the certain subject into the above-mentioned "warm type" or the like. In addition, in the third embodiment, the predetermined ratio is set to be 70% or 80%, but it can also be 50% or 60%. The number of types, the names of the types, and the like are not limited to the above examples, and can be set appropriately.

[0069] The association construction unit 4 may also classify each subject as described above based on the factor information from each subject, and analyze what degree of comfort or discomfort each subject will feel based on the comfort level from each subject. For example, when the factor information from a certain subject indicates coldness and the comfort level obtained from the certain subject together with the factor information is low, the association construction unit 4 may also analyze the certain subject as a "cold-sensitive type". In addition, the association construction unit 4 may also classify the certain subject as a "warm type" and a "cold-sensitive type", etc., and perform more detailed classification than when only the factor information is used. In addition, the association construction unit 4 may also classify each subject using not only the factor information from each subject but also the biological information from each subject. For example, when the biological information from a certain subject indicates low body temperature and the factor information obtained from the certain subject at the same time point as the biological information indicates "coldness", the association construction unit 4 may also analyze the certain subject as a "cold-sensitive type".

[0070] The association construction unit 4 in the third embodiment constructs association information that associates type information representing a type, the first biometric information representing the state of a body of one or more subjects of the type, and the emotion information corresponding to the first biometric information representing the state of a body of the one or more subjects. In addition, the emotion information corresponding to the first biometric information representing the state of a body of one or more subjects of the type may also be emotion information obtained from the one or more subjects at the same time point as the first biometric information. Alternatively, it may be information obtained by using the emotion information obtained from the one or more subjects at the same time point as the first biometric information and through the above-mentioned regression analysis or the above-mentioned average calculation processing. The emotion information corresponding to the first biometric information representing the state of a body of one or more subjects of the type is information that generalizes the emotions of one or more subjects of the type.

[0071] Hereinafter, an example will be described in which the first biometric information is the temporal variation of the body surface temperature. Assume that there are five subjects classified into a certain type, and the emotion information obtained from the five subjects at the same time point as the time point when the first biometric information indicating 0.5 [°C] was obtained indicates a high level of comfort. In this case, the numerical value indicating the high level of comfort or the result obtained by averaging the numerical value becomes an example of "emotion information corresponding to the temporal variation of the body surface temperature of the five subjects of the certain type", that is, "emotion information corresponding to the biometric information indicating the physical state of the five subjects of the certain type". In addition, when the relationship between the first biometric information and the emotion information obtained from the five subjects is formulated by regression analysis or the like using the respective values ​​of the first biometric information and the emotion information obtained from the five subjects, the value of the emotion information obtained by substituting 0.5 [°C] into the formula becomes an example of "emotion information corresponding to the biometric information indicating the physical state of the five subjects of the certain type". It is considered that the high comfort level generally indicates the emotion of the five subjects of the certain type at the time when the first biological information indicating 0.5 [° C.] is acquired.

[0072] The association construction unit 4 in Implementation Example 3 can also construct association information that associates identification information of each object person, type information indicating the type of each object person, the first biometric information indicating the physical state of each object person, and emotion information corresponding to the first biometric information indicating the physical state of each object person.

[0073] When the type information, the first biometric information, and the emotion information corresponding to the first biometric information are associated in the associated information, the estimation unit 6 extracts the emotion information associated with the first biometric information equal to the second biometric information acquired by the estimated biometric information acquisition unit 5. In the third embodiment, the type information is uniquely determined based on the value of the first biometric information. The estimation unit 6 may extract the type information associated with the first biometric information equal to the second biometric information, and estimate the type of the subject person who is the acquisition source of the second biometric information.

[0074] When the identification information, the type information, the first biometric information, and the emotion information corresponding to the first biometric information are associated in the associated information, the estimation unit 6 extracts the emotion information associated with the identification information acquired by the estimated biometric information acquisition unit 5 and the first biometric information equal to the second biometric information acquired by the estimated biometric information acquisition unit 5. The estimation unit 6 may extract the type information associated with the identification information and the first biometric information equal to the second biometric information, and estimate the type of the subject person who is the acquisition source of the second biometric information.

[0075] When the inference unit 6 extracts the type information, it instructs the output unit 7 to output the extracted type information to an external device. The output unit 7 outputs the type information to an external device. When the external device is the above-mentioned display device, the type information obtained from the output unit 7 is displayed. When the external device is the above-mentioned sound output device, the type information obtained from the output unit 7 is output by sound. When the external device is the above-mentioned terminal, the type information obtained from the output unit 7 is displayed or output by sound. Thus, the subject can know the type of the subject. In addition, the subject can also know the type of the subject by referring to the storage unit 1 using the terminal connected to the emotion inference device 100.

[0076] If the subject knows his / her own type, the subject may input type information indicating the type to an input device such as a keyboard or a touch panel (not shown) when acquiring the second biometric information. Furthermore, the estimation unit 6 may extract the subject's identification information and the type information, or emotion information associated with the type information, to estimate the subject's emotion.

[0077] Figure 5 2 is a flowchart illustrating the emotion estimation process according to Embodiment 3. Here, an example is given of a case where the biological information acquisition unit 2 and the emotion information acquisition unit 3 acquire identification information for identifying each object. Each process in step S41 to step S44 is the same as each process in step S21 to step S24 described above, so the description thereof is omitted.

[0078] In step S44, when the above period ends (step S44: "Yes"), in step S45, the association construction unit 4 performs cluster analysis to classify each subject into any of a plurality of types. In step S46, the association construction unit 4 constructs association information that associates type information, first biometric information indicating the physical state of one or more subjects of the type indicated by the type information, and emotion information indicating the emotions of the one or more subjects and corresponding to the first biometric information. Alternatively, in step S46, the association construction unit 4 constructs association information that associates the identification information of each subject, type information indicating the type of each subject, first biometric information indicating the physical state of each subject, and emotion information indicating the emotions of each subject and corresponding to the first biometric information. The processing in step S47 is the same as the processing in step S6, so the description is omitted. The processing in step S48 is the same as the processing in step S7 or step S27, so the description is omitted. The processing in step S49 is the same as the processing in step S8, so the description is omitted.

[0079] In step S50, the inference unit 6 extracts the emotion information and type information associated with the first biometric information that is equal to the second biometric information analyzed in step S49. Alternatively, the inference unit 6 extracts the emotion information and type information associated with the identification information of the above-mentioned arbitrary subject that identifies the acquisition source of the second biometric information and the first biometric information that is equal to the second biometric information. In step S51, the inference unit 6 instructs the output unit 7 to output the extracted emotion information and type information to an external device. After the processing of step S51, the emotion inference processing ends. In addition, when the emotion inference processing is executed again, the emotion inference device 100 starts the execution from the processing of step S48.

[0080] The effects of the emotion estimation device 200 according to the third embodiment will be described below. The prepared emotion information acquisition unit 3 in the third embodiment acquires, as emotion information, a comfort level indicating the degree of comfort of each of the plurality of subjects according to the evaluation grid method, and acquires factor information indicating the factors of the comfort level of each of the plurality of subjects. The prepared emotion information acquisition unit 3 acquires the comfort level according to the scoring method and acquires the factor information according to the free description method. The association construction unit 4 performs cluster analysis using the factor information from each of the plurality of subjects, and classifies each of the plurality of subjects into any of a plurality of categories. The association construction unit 4 constructs association information that associates category information indicating a category, first biometric information indicating the state of the body of one or more subjects among the plurality of subjects classified into the category, and emotion information indicating the emotion of the one or more subjects and corresponding to the first biometric information indicating the state of the body of the one or more subjects. The estimation unit 6 extracts the emotion information and category information that are associated with the first biometric information that is equal to the second biometric information indicating the state of the body of the arbitrary subject in the association information. The estimation unit 6 estimates the emotion of the arbitrary subject based on the extracted emotion information, and estimates the type of the arbitrary subject based on the extracted type information. Thus, the associated information becomes information indicating the correlation between the first biometric information and the emotion information that are specific to a type such as "warmth type" or "inner type". Therefore, the estimation unit 6 can estimate the emotion of the subject with high accuracy using the second biometric information according to the type of the subject.

[0081] The prepared emotion information acquisition unit 3 in the third embodiment acquires, from each of the plurality of subjects, a comfort level indicating the degree of comfort of each of the plurality of subjects as emotion information, and acquires factor information indicating the factors of the comfort level of each of the plurality of subjects, according to the evaluation grid method. The prepared emotion information acquisition unit 3 acquires the comfort level according to the scoring method, and acquires the factor information according to the free description method. The prepared biometric information acquisition unit 2 and the prepared emotion information acquisition unit 3 acquire identification information for identifying each of the plurality of subjects. The association construction unit 4 performs cluster analysis using the factor information from each of the plurality of subjects, and classifies each of the plurality of subjects into any of a plurality of categories. The association construction unit 4 constructs association information that associates the identification information of each of the plurality of subjects, the category information indicating the category of each of the plurality of subjects, the first biometric information indicating the state of the body of each of the plurality of subjects, and the emotion information indicating the emotion of each of the plurality of subjects and corresponding to the first biometric information indicating the state of the body of each of the plurality of subjects. The inference unit 6 extracts the emotion information and type information associated with the identification information of the arbitrary subject and the first biometric information equal to the second biometric information in the associated information. The inference unit 6 infers the emotion of the arbitrary subject based on the extracted emotion information, and infers the type of the arbitrary subject based on the extracted type information. Thus, the associated information becomes information indicating the correlation between the physical state and the emotion for each subject. Therefore, the inference unit 6 can infer the emotion for each subject by using the associated information. Therefore, the emotion inference device 100 can infer the emotion for each subject based on the type with high accuracy.

[0082] The association construction unit 4 in the third embodiment uses the first biological information from each subject and at least one of the comfort level from each subject, and the factor information from each subject to perform cluster analysis, and classify each subject into any of a plurality of types. Thus, the association construction unit 4 can classify each subject into types in detail. Therefore, the accuracy of the association information is further improved. Therefore, the accuracy of the emotion estimation device 100 in estimating emotions is improved.

[0083] The emotion estimation device 100 according to the third embodiment further includes an output unit 7 for communicating with an external device. The estimation unit 6 controls the output unit 7 in such a manner that the extracted emotion information and type information are sent to the external device. Thus, a user of an external device such as a terminal, a display device, or a sound output device can know the emotion and type of the arbitrary subject. In addition, when the user is the arbitrary subject, the arbitrary subject can know his or her own type.

[0084] Implementation method 4.

[0085] The emotion estimation device 100 according to Embodiment 4 estimates the emotion of the subject with higher accuracy. Embodiment 4 is described below. In addition, among the components in Embodiment 4, the components that are the same as the components in Embodiments 1 to 3 are given the same symbols as those in Embodiments 1 to 3. In addition, in the content of Embodiment 4, the description of the same contents as those in Embodiments 1 to 3 will be omitted unless there are special circumstances.

[0086] Function blocks included in the emotion estimation device 100 according to the fourth embodiment are Figure 1 The functional blocks in the emotion estimation device 100 according to the illustrated embodiments 1 to 3 are the same. In addition, the prepared emotion information acquisition unit 3 in the fourth embodiment acquires emotion information in the same manner as in the first embodiment described above.

[0087] Among the biological information indicating the physical state of a person, the difference in facial temperature from resting, the temperature difference between the cheek and the forehead, the value of the electrocardiogram index, or the value indicating the galvanic skin response is strongly related to the person's emotions. The biological information acquisition unit 2 in the fourth embodiment acquires the first biological information including at least one of the facial temperature, electrocardiogram data, and galvanic skin response data of each of a plurality of subjects. The biological information acquisition unit 2 detects the facial temperature using, for example, an infrared sensor. In addition, the biological information acquisition unit 2 acquires the electrocardiogram data or the galvanic skin response data using, for example, a wearable sensor.

[0088] The biological information acquisition unit 2 is prepared to calculate the first biological information including at least one of the difference from the resting facial temperature, the difference between the cheek and the forehead, the value of the electrocardiogram index, and the value indicating the galvanic skin response of each of the plurality of subjects based on the acquired first biological information. In addition, it is prepared that the biological information acquisition unit 2 acquires the resting facial temperature in advance when calculating the difference from the resting facial temperature.

[0089] The estimated biometric information acquisition unit 5 acquires at least one of the facial temperature, electrocardiogram data, and galvanic skin response data of any of the plurality of subjects. The estimated biometric information acquisition unit 5 detects the facial temperature using, for example, an infrared sensor. In addition, the estimated biometric information acquisition unit 5 acquires the electrocardiogram data or galvanic skin response data using, for example, a wearable sensor.

[0090] The estimated biometric information acquisition unit 5 calculates the second biometric information including at least one of the difference in facial temperature from the resting state, the temperature difference between the cheek and the forehead, the value of the electrocardiogram index, and the value indicating the skin electrical response of the arbitrary subject based on the acquired second biometric information.

[0091] As the above-mentioned electrocardiogram index, the high frequency band, low frequency band, ultra-low frequency band, or the standard deviation of the heartbeat interval are strongly related to emotions, so they are preferred. In addition, as the value representing the electrical response of the skin, the amount of sweat, the response amplitude, the occurrence frequency, the response time, or the recovery time are strongly related to emotions, so they are preferred.

[0092] Emotion Estimation Processing and Reference of Emotion Estimation Device 100 in Embodiment 4 Figure 3 to Figure 5 The emotion estimation processes according to Embodiments 1 to 3 described above are the same except for the following points.

[0093] The effect of the emotion estimation device 100 involved in the fourth embodiment is described below. The biological information acquisition unit 2 is prepared to acquire the first biological information including the facial temperature of each of the multiple subjects. The biological information acquisition unit 2 uses the facial temperature of each subject to calculate the first biological information including the difference between the facial temperature of each subject and the facial temperature of each subject when the subject is at rest, or the temperature difference between the cheek and the forehead of each subject. The association construction unit 4 constructs association information that associates the first biological information calculated by the biological information acquisition unit 2 and the emotion information corresponding to the first biological information. The estimated biological information acquisition unit 5 acquires the second biological information including the facial temperature of any subject among the above-mentioned multiple subjects. The estimated biological information acquisition unit 5 uses the facial temperature of the arbitrary subject to calculate the second biological information including the difference between the facial temperature of the arbitrary subject and the facial temperature of the arbitrary subject when the subject is at rest, or the temperature difference between the cheek and the forehead of the arbitrary subject. The estimation unit 6 extracts the emotion information associated with the first biometric information equal to the second biometric information calculated by the estimated biometric information acquisition unit 5, and estimates the emotion of the arbitrary subject based on the extracted emotion information. The difference from the facial temperature at rest and the temperature difference between the cheek and the forehead are strongly related to the emotion. Therefore, by using the difference or the temperature difference as the first biometric information and the second biometric information, the estimation accuracy of the emotion of the emotion estimation device 100 is improved.

[0094] The prepared biometric information acquisition unit 2 in the fourth embodiment acquires the first biometric information including the electrocardiogram data of each of the plurality of subjects. The prepared biometric information acquisition unit 2 uses the electrocardiogram data of each subject to calculate the first biometric information including the value of the electrocardiogram index of each subject. The association construction unit 4 constructs association information that associates the first biometric information calculated by the prepared biometric information acquisition unit 2 with the emotion information corresponding to the first biometric information. The estimated biometric information acquisition unit 5 acquires the second biometric information including the electrocardiogram data of any subject among the plurality of subjects. The estimated biometric information acquisition unit 5 uses the electrocardiogram data of the arbitrary subject to calculate the second biometric information including the value of the electrocardiogram index of the arbitrary subject. The estimation unit 6 extracts the emotion information associated with the first biometric information that is equal to the second biometric information calculated by the estimated biometric information acquisition unit 5, and estimates the emotion of the arbitrary subject based on the extracted emotion information. The electrocardiogram index is strongly correlated with the emotion. Therefore, by using the electrocardiogram index as the first biometric information and the second biometric information, the estimation accuracy of the emotion estimation device 100 is improved.

[0095] The prepared biometric information acquisition unit 2 in the fourth embodiment acquires the first biometric information including the skin galvanic response data of each of the plurality of subjects. The prepared biometric information acquisition unit 2 uses the skin galvanic response data of each subject to calculate the first biometric information including the value indicating the skin galvanic response of each subject. The association construction unit 4 constructs association information that associates the first biometric information calculated by the prepared biometric information acquisition unit 2 and the emotion information corresponding to the first biometric information. The estimated biometric information acquisition unit 5 acquires the second biometric information including the skin galvanic response data of any subject among the plurality of subjects. The estimated biometric information acquisition unit 5 uses the skin galvanic response data of the arbitrary subject to calculate the second biometric information including the value indicating the skin galvanic response of the arbitrary subject. The estimation unit 6 extracts the emotion information associated with the first biometric information that is equal to the second biometric information calculated by the estimated biometric information acquisition unit 5, and estimates the emotion of the arbitrary subject based on the extracted emotion information. The skin galvanic response is strongly correlated with the emotion. Therefore, by using the value indicating the galvanic skin response as the first biological information and the second biological information, the estimation accuracy of the emotion estimation device 100 is improved.

[0096] The biological information acquisition unit 2 and the estimated biological information acquisition unit 5 each have a wearable sensor. Thus, the biological information acquisition unit 2 can acquire the first biological information from each of the plurality of subjects on a daily basis. Therefore, the associated information becomes information that shows in detail the correlation between the physical state and the emotion of each subject in daily life. In addition, the estimated biological information acquisition unit 5 can acquire the second biological information of any of the above subjects in daily life. Thus, the emotion estimation device 100 can estimate the emotions of the subject in daily life with high accuracy.

[0097] Although the embodiment has been described above, the content of the present disclosure is not limited to the embodiment but includes conceivable equivalent scopes.

[0098] Explanation of symbols

[0099] 1: storage unit; 2: prepare biological information acquisition unit; 3: prepare emotion information acquisition unit; 4: association construction unit; 5: infer biological information acquisition unit; 6: inference unit; 7: output unit; 11: processor; 12: memory; 13: storage device; 14: first input interface circuit; 15: sensor; 16: second input interface circuit; 17: input device; 18: output interface circuit; 19: bus; 100: emotion inference device.

Claims

1. An emotion estimation device for estimating emotions of a plurality of subjects existing in an estimation space, wherein: The emotion estimation device comprises: a preparation biometric information acquisition unit that acquires and analyzes first biometric information indicating a physical state of each of the plurality of subjects present in the preparation space; a preparation emotion information acquisition unit that acquires emotion information indicating the emotion of each of the plurality of subjects present in the preparation space; a correlation constructing unit that constructs correlation information that correlates the first biometric information and the emotion information corresponding to the first biometric information; an estimated biometric information acquisition unit that acquires and analyzes second biometric information indicating a body state of an arbitrary subject among the plurality of subjects that is present in the estimated space; as well as an inference unit, extracting the emotion information associated with the first biometric information equal to the second biometric information in the associated information, and inferring the emotion of the arbitrary subject existing in the inference space based on the extracted emotion information, The prepared emotion information acquisition unit acquires, from each of the plurality of subjects, a comfort level indicating a degree of comfort of each of the plurality of subjects as the emotion information, and acquires factor information indicating a factor of the comfort level of each of the plurality of subjects, based on an evaluation grid method. The association building unit performs cluster analysis using the factor information from each of the plurality of subjects to classify each of the plurality of subjects into any one of a plurality of categories. The association construction unit constructs the association information by associating the type information indicating the type, the first biometric information indicating the physical state of one or more subjects among the plurality of subjects classified as the type, and the emotion information indicating the emotion of the one or more subjects and corresponding to the first biometric information indicating the physical state of the one or more subjects, The inference unit extracts the emotion information and the type information associated with the first biometric information that is equal to the second biometric information representing the physical state of the arbitrary object person in the associated information, infers the emotion of the arbitrary object person based on the extracted emotion information, and infers the type of the arbitrary object person based on the extracted type information.

2. An emotion estimation device for estimating emotions of a plurality of subjects existing in an estimation space, wherein: The emotion estimation device comprises: a preparation biometric information acquisition unit that acquires and analyzes first biometric information indicating a physical state of each of the plurality of subjects present in the preparation space; a preparation emotion information acquisition unit that acquires emotion information indicating the emotion of each of the plurality of subjects present in the preparation space; a correlation constructing unit that constructs correlation information that correlates the first biometric information and the emotion information corresponding to the first biometric information; an estimated biometric information acquisition unit that acquires and analyzes second biometric information indicating a body state of an arbitrary subject among the plurality of subjects that is present in the estimated space; as well as an inference unit, extracting the emotion information associated with the first biometric information equal to the second biometric information in the associated information, and inferring the emotion of the arbitrary subject existing in the inference space based on the extracted emotion information, The prepared emotion information acquisition unit acquires, from each of the plurality of subjects, a comfort level indicating a degree of comfort of each of the plurality of subjects as the emotion information, and acquires factor information indicating a factor of the comfort level of each of the plurality of subjects, based on an evaluation grid method. The preparation biological information acquisition unit and the preparation emotion information acquisition unit acquire identification information for identifying each of the plurality of subjects, The association building unit performs cluster analysis using the factor information from each of the plurality of subjects to classify each of the plurality of subjects into any one of a plurality of categories. The association constructing unit constructs the association information by associating the identification information of each of the plurality of subjects, the type information indicating the type of each of the plurality of subjects, the first biometric information indicating the physical state of each of the plurality of subjects, and the emotion information indicating the emotion of each of the plurality of subjects and corresponding to the first biometric information indicating the physical state of each of the plurality of subjects, The inference unit extracts the emotion information and the type information associated with the identification information of the arbitrary object person and the first biometric information equal to the second biometric information in the associated information, infers the emotion of the arbitrary object person based on the extracted emotion information, and infers the type of the arbitrary object person based on the extracted type information.

3. The emotion estimation device according to claim 1 or 2, wherein: The association construction unit constructs the association information using the emotion information obtained from each of the multiple subjects by the preparation emotion information acquisition unit and the first biometric information obtained from each of the multiple subjects by the preparation biometric information acquisition unit within a predetermined time before and after the time point when the preparation emotion information acquisition unit obtained the emotion information from each of the multiple subjects.

4. The emotion estimation device according to claim 1 or 2, wherein: The preparation emotion information acquisition unit acquires the emotion information indicating the emotion of each of the plurality of subjects at each of the plurality of time points within a predetermined period from each of the plurality of subjects present in the preparation space. The association construction unit constructs the association information using the first biometric information acquired from each of the plurality of subjects by the preparation biometric information acquisition unit at each of the plurality of time points and the emotion information acquired from each of the plurality of subjects by the preparation emotion information acquisition unit at each of the plurality of time points.

5. The emotion estimation device according to claim 4, wherein: The period is a period of more than 2 days, The preparation emotion information acquisition unit acquires the emotion information from each of the plurality of subjects at one or more predetermined times or at one or more random times on each date.

6. The emotion estimation device according to claim 5, wherein: The period is more than 5 days.

7. The emotion estimation device according to claim 1 or 2, wherein: The preparation emotion information acquisition unit acquires the comfort level according to a scoring method, The preparation emotion information acquisition unit acquires the factor information according to a free description method.

8. The emotion estimation device according to claim 1 or 2, wherein: The association construction unit uses the first biometric information from each of the multiple subjects, at least one of the comfort levels from each of the multiple subjects, and the factor information from each of the multiple subjects to perform cluster analysis, and classify each of the multiple subjects into any of the multiple types.

9. The emotion estimation device according to claim 1 or 2, wherein: The emotion estimation device further includes an output unit that communicates with an external device. The estimation unit controls the output unit so as to transmit the extracted emotion information and the category information to the external device.

10. The emotion estimation device according to claim 1 or 2, wherein: The emotion estimation device further includes an output unit that communicates with an external device. The estimation unit controls the output unit so as to transmit the extracted emotion information to the external device.

11. The emotion estimation device according to claim 1 or 2, wherein: The preparation biometric information acquisition unit acquires the first biometric information including the facial temperature of each of the plurality of subjects, and uses the facial temperature of each of the plurality of subjects to calculate the first biometric information including the difference between the facial temperature of each of the plurality of subjects and the facial temperature of each of the plurality of subjects when the plurality of subjects are at rest, or the temperature difference between the cheek and the forehead of each of the plurality of subjects, the association constructing unit constructs the association information by associating the first biometric information calculated by the preparation biometric information acquiring unit with the emotion information corresponding to the first biometric information calculated by the preparation biometric information acquiring unit, The estimated biometric information acquisition unit acquires the second biometric information including the facial temperature of the arbitrary subject, and uses the facial temperature of the arbitrary subject to calculate the second biometric information including the difference between the facial temperature of the arbitrary subject and the facial temperature of the arbitrary subject when the arbitrary subject is at rest, or the temperature difference between the cheek and forehead of the arbitrary subject. The estimation unit extracts the emotion information associated with the first biometric information that is equal to the second biometric information calculated by the estimated biometric information acquisition unit, and estimates the emotion of the arbitrary subject based on the extracted emotion information.

12. The emotion estimation device according to claim 11, wherein: The prepared biological information acquisition unit and the estimated biological information acquisition unit each include an infrared sensor. The infrared sensor detects the facial temperature.

13. The emotion estimation device according to claim 1 or 2, wherein: The preparation biological information acquisition unit acquires the first biological information including electrocardiogram data of each of the plurality of subjects, and calculates the first biological information including the value of the electrocardiogram index of each of the plurality of subjects using the electrocardiogram data of each of the plurality of subjects, the association constructing unit constructs the association information by associating the first biometric information calculated by the preparation biometric information acquiring unit with the emotion information corresponding to the first biometric information calculated by the preparation biometric information acquiring unit, The estimated biometric information acquisition unit acquires the second biometric information including electrocardiogram data of the arbitrary subject, and calculates the second biometric information including a value of an electrocardiogram index of the arbitrary subject using the electrocardiogram data of the arbitrary subject. The estimation unit extracts the emotion information associated with the first biometric information that is equal to the second biometric information calculated by the estimated biometric information acquisition unit, and estimates the emotion of the arbitrary subject based on the extracted emotion information.

14. The emotion estimation device according to claim 1 or 2, wherein: The preparation biometric information acquisition unit acquires the first biometric information including skin galvanic response data of each of the plurality of subjects, and uses the skin galvanic response data of each of the plurality of subjects to calculate the first biometric information including a value indicating the skin galvanic response of each of the plurality of subjects, the association constructing unit constructs the association information by associating the first biometric information calculated by the preparation biometric information acquiring unit with the emotion information corresponding to the first biometric information calculated by the preparation biometric information acquiring unit, The estimated biometric information acquisition unit acquires the second biometric information including the skin galvanic response data of the arbitrary subject, and uses the skin galvanic response data of the arbitrary subject to calculate the second biometric information including a value indicating the skin galvanic response of the arbitrary subject, The estimation unit extracts the emotion information associated with the first biometric information that is equal to the second biometric information calculated by the estimated biometric information acquisition unit, and estimates the emotion of the arbitrary subject based on the extracted emotion information.

15. The emotion estimation device according to claim 13, wherein: The prepared biological information acquisition unit and the estimated biological information acquisition unit each include a wearable sensor.

16. An emotion estimation method, which is executed by an emotion estimation device for estimating emotions of a plurality of subjects existing in an estimation space, wherein: The emotion inference method comprises: a preparation biometric information acquisition step of acquiring and analyzing first biometric information indicating a physical state of each of the plurality of subjects present in the preparation space; a preparation emotion information acquisition step of acquiring emotion information indicating the emotion of each of the plurality of subjects existing in the preparation space; a correlation constructing step of constructing correlation information that correlates the first biometric information and the emotion information corresponding to the first biometric information; an estimated biometric information acquisition step of acquiring and analyzing second biometric information indicating a body state of an arbitrary subject among the plurality of subjects existing in the estimated space; and an inference step of extracting the emotion information associated with the first biometric information equal to the second biometric information in the associated information, and inferring the emotion of the arbitrary subject existing in the inference space based on the extracted emotion information, In the step of obtaining the prepared emotion information, a comfort level indicating the degree of comfort of each of the plurality of subjects is obtained as the emotion information according to the evaluation grid method, and factor information indicating the factors of the comfort level of each of the plurality of subjects is obtained. In the association building step, cluster analysis is performed using the factor information from each of the plurality of subjects to classify each of the plurality of subjects into any of a plurality of categories. In the association construction step, the association information is constructed by associating type information indicating the type, first biometric information indicating the physical state of one or more subjects among the plurality of subjects classified as the type, and emotion information indicating the emotion of the one or more subjects and corresponding to the first biometric information indicating the physical state of the one or more subjects, In the inference step, the emotion information and the type information associated with the first biometric information equal to the second biometric information representing the physical state of the arbitrary object person are extracted from the associated information, the emotion of the arbitrary object person is inferred based on the extracted emotion information, and the type of the arbitrary object person is inferred based on the extracted type information.

17. An emotion estimation method, which is executed by an emotion estimation device for estimating emotions of a plurality of subjects existing in an estimation space, wherein: The emotion inference method comprises: a preparation biometric information acquisition step of acquiring and analyzing first biometric information indicating a physical state of each of the plurality of subjects present in the preparation space; a preparation emotion information acquisition step of acquiring emotion information indicating the emotion of each of the plurality of subjects existing in the preparation space; a correlation constructing step of constructing correlation information that correlates the first biometric information and the emotion information corresponding to the first biometric information; an estimated biometric information acquisition step of acquiring and analyzing second biometric information indicating a body state of an arbitrary subject among the plurality of subjects existing in the estimated space; and an inference step of extracting the emotion information associated with the first biometric information equal to the second biometric information in the associated information, and inferring the emotion of the arbitrary subject existing in the inference space based on the extracted emotion information, In the step of obtaining the prepared emotion information, a comfort level indicating the degree of comfort of each of the plurality of subjects is obtained as the emotion information according to the evaluation grid method, and factor information indicating the factors of the comfort level of each of the plurality of subjects is obtained. In the preparation biological information acquisition step and the preparation emotion information acquisition step, identification information for identifying each of the plurality of subjects is acquired, In the association building step, cluster analysis is performed using the factor information from each of the plurality of subjects to classify each of the plurality of subjects into any of a plurality of categories. In the association building step, the association information is built by associating the identification information of each of the plurality of subjects, the type information indicating the type of each of the plurality of subjects, the first biometric information indicating the physical state of each of the plurality of subjects, and the emotion information indicating the emotion of each of the plurality of subjects and corresponding to the first biometric information indicating the physical state of each of the plurality of subjects, In the inference step, the emotion information and the type information associated with the identification information of the arbitrary object person and the first biometric information equal to the second biometric information are extracted from the associated information, the emotion of the arbitrary object person is inferred based on the extracted emotion information, and the type of the arbitrary object person is inferred based on the extracted type information.

Images