JP7848550B2 - Estimation method, estimation device, and estimation program - Google Patents

Description

This invention relates to an estimation method, estimation apparatus, and estimation program for estimating either the relationship between subjective evaluation and the evaluator's state, or the properties of the subject.

It is generally recognized empirically that subjective evaluations of an object change under various conditions, even if the object itself does not change; for example, food tastes better when one is in a good mood. Regarding techniques for estimating emotions in relation to a person's state, for example, Patent Document 1 describes a method for estimating a user's emotions based on their voice. However, objective analysis of the relationship between the evaluator's state, including their emotions, and their subjective evaluation has not been conducted until now.

Japanese Patent Publication No. 2021-012285

The inventors analyzed changes in subjective evaluations under various evaluation conditions and examined the relationship between subjective evaluations and evaluation conditions. They found a correlation between subjective evaluations and the evaluator's state of mind.

Furthermore, by analyzing the relationship between the above-mentioned correlation and the nature of the object of evaluation, the inventors found that a certain relationship exists between the correlation between subjective evaluation and the evaluator's state and the nature of the object of evaluation.

Based on the above research findings, the present invention aims to provide a novel technique for estimating either the relationship between subjective evaluation and the evaluator's state, or the properties of the subject, by utilizing the aforementioned relationship.

To solve the above problems, the present invention provides an estimation method for estimating information regarding changes in subjective evaluation of an object according to the state of the evaluator, based on the properties of the object,
A step of acquiring relationship information relating to the relationship between state information relating to the state of the evaluator and evaluation information indicating the evaluator's subjective evaluation of the subject, and property relationship information indicating the relationship between the property information indicating the property of the subject,
The steps include obtaining property information about the object to be estimated,
The process includes the step of estimating the relational information from the property information of the object to be estimated, based on the aforementioned property relational information.

This configuration makes it possible to estimate the relationship between subjective evaluation and the evaluator's state when evaluating a specific target, based on its properties. This allows, for example, the estimation of changes in subjective evaluation related to the evaluator's state for a particular product, which can then be used to inform product marketing activities such as suggesting effective usage methods or developing scientifically-based catchphrases.

To solve the above problems, the present invention provides an estimation method for estimating the properties of an object such that the subjective evaluation of the object shows a specific change depending on the state of the evaluator,
A step of acquiring relationship information relating to the relationship between state information relating to the state of the evaluator and evaluation information indicating the evaluator's subjective evaluation of the subject, and property relationship information indicating the relationship between the property information indicating the property of the subject,
A step of specifying relationship information that indicates the desired relationship,
The process includes the step of estimating the property information of an object that corresponds to the specified relationship information, based on the property relationship information.

This configuration allows for the estimation of the properties of an object that exhibit a desired relationship. This makes it possible, for example, to estimate the properties of an object that show a change in subjective evaluation expected under specific conditions, which can be useful in product design. More specifically, for example, to design a product that exhibits a certain level of positive correlation between the evaluator's level of arousal and the evaluation value indicating luxury, it is conceivable to estimate the properties necessary for this correlation and then design the product with these estimated properties as the target.

In a preferred embodiment of the present invention, the property information is a value representing the properties of the object.

In a preferred embodiment of the present invention, the object is an object, and the property information is the physical property value of the object.

In a preferred embodiment of the present invention, the aforementioned physical property values are those that represent properties perceptible to human touch.

In a preferred embodiment of the present invention, the state information includes a state value indicating the mental state of the evaluator at the time of evaluation.

In a preferred embodiment of the present invention, the state value includes an emotional valence indicating the evaluator's feelings.

In a preferred embodiment of the present invention, the state value includes the evaluator's level of alertness.

In a preferred embodiment of the present invention, the state value is represented by a combination of the emotional valence, which indicates the evaluator's emotions, and the evaluator's level of arousal.

In a preferred embodiment of the present invention, the relational information includes a value relating to the magnitude of the correlation between the evaluation information and the state information.

In a preferred embodiment of the present invention, the relational information includes a value relating to the number or percentage of evaluators showing a correlation between the evaluation information and the status information.

In a preferred embodiment of the present invention, the evaluation information is a subjective evaluation value of the subject by the evaluator, using evaluation terms as indicators.

In a preferred embodiment of the present invention, the evaluation value is a sensitivity evaluation value indicating the strength with which the evaluator perceives the sensation represented by the evaluation term.

The present invention is an estimation device for estimating information regarding changes in subjective evaluation of an object according to the state of the evaluator, based on the properties of the object.
Means for acquiring property relationship information, which indicates the relationship between status information relating to the state of the evaluator and evaluation information indicating the evaluator's subjective evaluation of the subject, and property information indicating the properties of the subject.
Regarding the object to be estimated, a means for obtaining property information,
The system includes means for estimating the relational information from the property information of the object to be estimated, based on the aforementioned property relational information.

The present invention is an estimation device for estimating the properties of an object such that the subjective evaluation of the object shows a specific change depending on the state of the evaluator,
Means for acquiring property relationship information, which indicates the relationship between status information relating to the state of the evaluator and evaluation information indicating the evaluator's subjective evaluation of the subject, and property information indicating the properties of the subject.
Means for specifying relational information that indicates a desired relationship,
The system includes means for estimating the property information of an object that corresponds to the specified relationship information, based on the aforementioned property relationship information.

The present invention is an estimation program for estimating information regarding changes in subjective evaluation of an object according to the state of the evaluator, based on the properties of the object, and is provided for use by a computer.
A step of acquiring relationship information relating to the relationship between state information relating to the state of the evaluator and evaluation information indicating the evaluator's subjective evaluation of the subject, and property relationship information indicating the relationship between the property information indicating the property of the subject,
The steps include obtaining property information about the object to be estimated,
Based on the aforementioned property relationship information, the system performs the step of estimating the relationship information from the property information of the object to be estimated.

The present invention is an estimation program for estimating the properties of an object such that the subjective evaluation of the object shows a specific change depending on the state of the evaluator, and is provided for a computer.
A step of acquiring relationship information relating to the relationship between state information relating to the state of the evaluator and evaluation information indicating the evaluator's subjective evaluation of the subject, and property relationship information indicating the relationship between the property information indicating the property of the subject,
A step of specifying relationship information that indicates the desired relationship,
Based on the aforementioned property relationship information, the system performs the step of estimating the property information of the target corresponding to the specified relationship information.

According to the present invention, it is possible to provide a novel technique for estimating either the relationship between a person's subjective evaluation and the evaluator's state, or the properties of the object.

This figure shows an example of the functional configuration of an estimation device that estimates information regarding changes in subjective evaluations according to the evaluator's state, based on the properties of the object being estimated. This figure shows an example of property relationship information in the estimation device of this embodiment. This figure shows an example of an estimation method that estimates information about changes in subjective evaluations in response to the evaluator's state, based on the properties of the object being estimated. This figure shows an example of the functional configuration of an estimation device that estimates the properties of an object such that the subjective evaluation of the object shows a specific change depending on the evaluator's state. This figure shows an example of a procedure for an estimation method that estimates the properties of an object in which subjective evaluations of the object show specific changes depending on the evaluator's state. This figure shows the combinations that showed a correlation between the correlation coefficient between emotional valence and evaluation value and the physical properties of the object in the example. This figure shows the combinations of the number of subjects who showed a correlation between emotional valence and evaluation value, and the physical properties of the object, in the example.

<Definition>
This invention relates to a technique for making estimations based on the relationship between the evaluator's state and the evaluator's subjective evaluation of the object of evaluation, and the nature of the object of evaluation. Here, "object" refers to something or a phenomenon that is the subject of evaluation, and in particular, it refers to an event that has stimuli that can be perceived by humans. In this invention, "object" includes, for example, objects, sounds, light, chemical substances, etc. In this embodiment, objects are mainly assumed to be objects and may be expressed as "objects," but it is also possible to replace them with things or events other than objects.

Furthermore, the properties of the object include various properties perceived by humans, such as mechanical, thermal, electrical, magnetic, and optical properties, as well as properties related to the object, such as sound properties and chemical properties. In particular, in this embodiment, an object is assumed as the object, and the physical properties of the object are analyzed as the properties of the object. In this invention, information indicating the properties of the object is called property information, and in this embodiment, information representing physical properties, such as physical property values, is used.

Furthermore, in this invention, "subjective evaluation" refers to the evaluator's subjective assessment of the subject. Hereafter, the information representing this evaluation will be referred to as evaluation information. While the evaluation method is not limited, it is generally assumed that evaluation values using evaluation terms as indicators will be used as evaluation information.

Here, "evaluative terms" refer to words that express subjective evaluations. These terms include, for example, terms expressing emotions such as "like," "dislike," "fun," "sad," and "surprised" (emotional terms); terms expressing abstract impressions of objects such as "luxury," "attractive," "special," and "gentle" (impressional terms); and terms expressing sensations directly evoked by objects themselves, such as "hard," "soft," "moist," and "dry" (sensory terms). In the following, emotional and impressional terms will be referred to as "higher-order terms," and sensory terms as "lower-order terms."

Furthermore, evaluation values based on evaluation terms refer to values obtained through the evaluator's subjective evaluation, which considers how well the object fits the impression represented by each evaluation term.
Preferably, the evaluation value is a sensitivity value indicating the strength of the sensation that the evaluator feels, as represented by the evaluation term. The evaluation value can be expressed as a number or integer between 0 and 100, for example, but the minimum and maximum values can be arbitrarily defined. Furthermore, the evaluator's subjective evaluation may be obtained as a binary value indicating whether or not the evaluation term is appropriate for the object, and this may also be used as the evaluation value.

Furthermore, "evaluator's state" refers to the evaluator's own state, typically their mental state. Other examples include biological attributes such as the evaluator's age and gender. In this invention, such information indicating the evaluator's state is called state information. In this embodiment, a state value, which is a value indicating the state, is used as state information.

Here, the evaluator's mental state broadly encompasses the state of a person's mind, but typically refers to the evaluator's emotions. The state value is the emotional value, which is a state value related to emotions. In particular, in this embodiment, referencing Russell's Circle of Emotions model, which expresses emotions on two axes—emotional valence (a state value indicating the degree of pleasure-displeasure) and arousal level (a state value indicating the degree of arousal)—the evaluator's emotions are assessed by a combination of these two state values. That is, the emotional value in this embodiment is a combination of emotional valence and arousal level. However, the method of expressing emotions is not limited to the combination of emotional valence and arousal level; any value related to emotions can be used as the emotional value. For example, only emotional valence or arousal level may be used as the emotional value, or Plutchik's Wheel of Emotions may be used. Furthermore, the strength and degree of various types of emotions, including joy, anger, sadness, and happiness, can also be used. Specifically, the intensity of emotions expressed in language, such as joy and sadness, and fatigue levels can also be used as emotional values.

The emotional value may be estimated based on biometric indicators such as electroencephalogram (EEG), skin potential, skin temperature, and heart rate variability, or it may be obtained through the evaluator's self-reporting. In this embodiment, both arousal level and emotional valence are used as the emotional value, and both are expressed by the evaluator themselves as integers between 0 and 100, although the minimum and maximum values can be arbitrarily defined.

In this invention, these indicators are used to obtain and estimate the relationship between relational information concerning the relationship between state information and evaluation information, and the property information of the object being evaluated. Relational information refers to information indicating the relationship between state information and evaluation information. For example, if state information is expressed as a function of evaluation information, that function, or combinations of types of state information and evaluation information that are related to each other, are used as relational information. Here, it is preferable that this relational information is linked to the property information of the object being evaluated.

The relationship information may show the relationship between status information and evaluation information for a single evaluator, or it may be a composite of the relationships between status information and evaluation information for multiple evaluators. Specifically, for example, it may be the correlation coefficient or p-value between status information and evaluation information for a single evaluator. Alternatively, it may be the average value of the correlation coefficients between status information and evaluation information for multiple evaluators, or the number of evaluators who showed a certain level of correlation between status information and evaluation information.

More specifically, for example, if correlation analysis reveals a correlation between emotional valence (an emotional value representing "pleasant-unpleasant") and the evaluation value using the evaluation term "pleasant," then information such as the combination of state information types and evaluation information types (emotional valence and the evaluation value of "pleasant"), the correlation coefficient and p-value for each evaluator in that combination, the average correlation coefficient among multiple evaluators, and the number or percentage of evaluators who showed a correlation above a certain level are assumed to be related information and linked to the physical properties of the object. Of these, the correlation coefficient and p-value can be used as values relating to the magnitude of the correlation. Note that "showing a correlation above a certain level" for a particular evaluator refers to, for example, when the same evaluator evaluates the same object multiple times under different conditions, the absolute value of the correlation coefficient between the state information and the evaluation information exceeds a threshold (e.g., 0.4), or the p-value is below a threshold.

Furthermore, the present invention estimates relational information or property information based on property relational information that shows the relationship between the above-mentioned relational information and the property information of the object. Property relational information is information that represents the relationship between specific relational information and specific property information. For example, if relational information is expressed as a function of property information, that function, or the types of relational information (combinations of state information types and evaluation value types) and combinations of state information types that are related to each other, are used as property relational information.

In this embodiment, the relational information stored includes, for each combination of state information type and evaluation value type, the average correlation coefficient of multiple evaluators, the number of evaluators who showed a correlation of a certain level or higher in that combination, and the total number of evaluators.
Then, as property relationship information, for each of the above relationship information (for each combination of state information and evaluation value types), the correlation coefficient between the physical property value (property information) and the relationship information is stored. More specifically, for each type of physical property value and each type of relationship information, the correlation coefficient between the physical property value and the average correlation coefficient among multiple evaluators, and the correlation coefficient between the physical property value and the number or percentage of evaluators who showed a correlation of a certain level or higher are stored as property relationship information.

<Implementation>
Embodiments of the present invention will be described below with reference to the drawings. Note that the embodiments shown below are examples of the present invention, and the present invention is not limited to these embodiments; various configurations can be adopted.

For example, while this embodiment primarily describes the configuration and operation of the estimation device, similar configurations, computer programs, etc., can achieve similar effects. Furthermore, the program may be stored on a recording medium. Using this recording medium, the program can be installed on a computer, for example. Here, the recording medium storing the program may be a non-transient recording medium such as a CD-ROM.

[1] Estimation of relational information First, we will explain the estimation device and estimation method for estimating relational information from property information. In the present invention, the object for which we want to estimate relational information is called the estimation target, and relational information is estimated from the property information of the estimation target based on property relational information that shows the relationship between property information and relational information.

Figure 1 shows an example of the functional configuration of the estimation device 1, which estimates information regarding changes in subjective evaluation according to the evaluator's state based on the properties of the target of estimation. The estimation device 1 can be a general computer device such as a server, equipped with various input/output devices including a CPU (Central Processing Unit) and GPU (Graphics Processing Unit), main memory such as RAM (Random Access Memory), auxiliary storage such as HDD (Hard Disk Drive), SSD (Solid State Drive), and flash memory, and means for connecting to a network. By installing a program for operating the computer as each means of the estimation device 1 in this invention into a storage device, any computer can function as the estimation device 1 of this invention.

The estimation device 1 comprises a property information acquisition means 11, a property relationship information acquisition means 12, an estimation means 13, and a database DB. These means do not need to be located in a single computer; multiple computers may cooperate to function as the estimation device 1 of this invention.

As described above, this invention utilizes property-relational information that shows the relationship between state information (indicating the evaluator's state) and evaluation information, and property information (indicating the object's properties). In the estimation device 1 of this embodiment, property-relational information is pre-registered in a database DB. Figure 2 shows an example of property-relational information stored in the database DB. For example, it is conceivable to store property-relational information showing the relationship between the average correlation coefficient (relationship information) between the state (state information) and evaluation value (evaluation information) for multiple evaluators, and the physical properties (property information) of the object. Alternatively, property-relational information between the number of evaluators showing a correlation between state and evaluation value, and the physical properties of the object, may also be stored. As shown in Figure 2, property-relational information is stored for each type of physical property, state, and evaluation value. In addition, the relationships for each type of state and evaluation value may be stored for each subject along with the physical properties of the object. Relationship information may also be stored.

The property information acquisition means 11 acquires property information about the object to be estimated. For example, if physical property values are used as property information, the physical property values of the object to be estimated should be acquired. The acquisition method can be arbitrarily determined; it may be specified numerically, or the property information acquisition means 11 may acquire physical property values measured by a measuring device connected to the estimation device 1. Alternatively, property information for multiple objects may be pre-registered in a database DB, and the property information of the object to be estimated may be acquired by accepting the selection of the object to be estimated from the registered objects.

The property relationship information acquisition means 12 acquires property relationship information from the database DB according to the type of property information acquired by the property information acquisition means 11. Here, the system may accept a specification of a combination of state information type and evaluation information type as the relationship to be estimated, and acquire property relationship information for the acquired property information and the relationship information of the specified combination. Furthermore, if the property information acquisition means 11 acquires multiple types of physical property values as property information, the system may also accept a specification of the types of property information.

The estimation means 13 estimates relational information from the property information acquisition means 11, based on the property relational information acquired by the property relational information acquisition means 12. For example, if property relational information is acquired between property information and the correlation coefficient between specific state information and evaluation information, the correlation coefficient between the specific state information and evaluation information can be estimated as relational information based on the property information. If there are multiple relational information items related to the acquired property information, estimation may be performed for each of them.

Next, referring to Figure 3, we will explain the estimation procedure using the estimation device 1 with a specific example. First, in step S11, the property information acquisition means 11 acquires property information of the object to be estimated. Here, we assume that the property information acquired is a physical property value representing the static friction force on the surface of the object.

Next, in step S12, the property relationship information acquisition means 12 identifies the property information obtained in step S11, in this case, the property relationship information showing the relationship with static friction force. For example, as shown in Figure 2, if property relationship information showing the relationship between static friction force and the average correlation coefficient between emotional valence and the evaluation value using the evaluation term "pleasant" is stored in the database DB, the property relationship information acquisition means 12 can acquire this property relationship information.

In the example shown in Figure 2, other relationship information relating to static friction is also registered. Therefore, in this case, multiple relationship information items can be estimated based on the registered relationship information. In such cases, the relationship information acquisition means 12 may accept a selection of the desired relationship information from the user and acquire the corresponding relationship information. Alternatively, all relationship information related to static friction may be acquired.

Then, in step S13, the estimation means 13 estimates relational information from the property information acquisition means 11, based on the property relational information acquired by the property relational information acquisition means 12. In this example, in step S12, property relational information showing the relationship between static friction force and the average correlation coefficient between emotional valence and the evaluation value using the evaluation term "pleasant" is acquired. Using the static friction force to be estimated, acquired in step S11, as input, the average correlation coefficient between emotional valence and the evaluation value using the evaluation term "pleasant" is estimated. If the property relational information acquisition means 12 acquires multiple property relational information, all the relational information that can be estimated from them may be estimated.

In this way, relational information showing the relationship between state information and evaluation information can be estimated from the property information of a specific target. This allows us to estimate what kind of relationship exists between the evaluator's state and the evaluation information when the target is evaluated. For example, by estimating relational information using a specific product as the target, it is possible to estimate how consumers will evaluate the product depending on their state at the time of use. This makes it possible to make scientifically based appeals in product promotions tailored to usage scenarios.

[2] Estimation of property information Next, we will describe the estimation device and estimation method for estimating property information from relational information. Note that the parts that are common with estimation device 1 will be omitted from the explanation. In the present invention, property relational information that shows the relationship between property information and relational information is used to estimate property information that exhibits a desired relationship from relational information that shows that relationship.

Figure 4 shows an example of the functional configuration of estimation device 2, which estimates the properties of an object in which the subjective evaluation of the object shows a specific change depending on the evaluator's state. Similar to estimation device 1, estimation device 2 can utilize a general computer device such as a server, equipped with arithmetic units such as a CPU (Central Processing Unit) and GPU (Graphics Processing Unit), main memory such as RAM (Random Access Memory), auxiliary storage such as an HDD (Hard Disk Drive), SSD (Solid State Drive), and flash memory, and various input/output devices including network connection means. By installing a program for operating the computer as each means of estimation device 2 in this invention into a storage device, any computer can function as estimation device 2 of this invention.

The estimation device 2 comprises a relational information designation means 21, a property relational information acquisition means 22, an estimation means 23, and a database DB. These means do not need to be provided on a single computer; multiple computers may cooperate to function as the estimation device 2 of the present invention. The information stored in the database DB is the same as that of the estimation device 1 described in [1], so its explanation is omitted.

The relationship information specification means 21 specifies relationship information that indicates the desired relationship. For example, it first accepts a selection from the user of a combination of expected state information and evaluation information, then accepts input of target relationship information, and specifies the input relationship information. Specifically, it is envisioned that, for example, a combination of emotional valence and an evaluation value using the evaluation term "pleasant" might be specified, and the relationship information might be, for example, a correlation coefficient value. Furthermore, the relationship information specification means 21 may also accept and specify a selection of the type of property information to be estimated.

The property relationship information acquisition means 22 acquires property relationship information from the database DB according to the type of relationship information specified by the relationship information specification means 21. Here, as with the estimation device 1, if there are multiple property relationship information entries related to the specified relationship information, the system may accept a selection of the property information to be estimated, or it may acquire multiple property relationship information entries related to the specified relationship information.

The estimation means 23 estimates property information based on the property relationship information acquired by the property relationship information acquisition means 22, and from the relationship information specified by the relationship information specification means 21. For example, if property relationship information is acquired between property information and the correlation coefficient between specific state information and evaluation information, the corresponding property information can be estimated based on the correlation coefficient. If there are multiple property information items related to the specified relationship information, estimation may be performed for each of them.

Next, referring to Figure 5, a specific example of the estimation procedure using the estimation device 2 will be explained. First, in step S21, the relation information specification means 21 specifies relation information indicating the desired relationship. Here, we assume that the average correlation coefficient between emotional valence and evaluation values based on the evaluation term "pleasant" is specified as the relation information. Furthermore, in this embodiment, the relation information specification means 21 also accepts the user's selection in step S21 and specifies the type of property information indicating this relation information. Here, we assume that a physical property value representing static friction force is selected as the type of property information.

Next, in step S22, the property relationship information acquisition means 22 identifies the property relationship information corresponding to the types of relationship information and property information specified in step S21. For example, as shown in Figure 2, if property relationship information showing the relationship between static friction force and the average correlation coefficient between emotional valence and evaluation values based on the evaluation term "pleasant" is stored in the database DB, the property relationship information acquisition means 22 can acquire this property relationship information according to the specification in step S21.

Then, in step S23, the estimation means 23 estimates the property information based on the property relationship information acquired by the property relationship information acquisition means 22, and from the relationship information specified by the relationship information specification means 21. In this example, the static friction force of an object exhibiting the specified average correlation coefficient is estimated using the desired numerical value for the average correlation coefficient between the emotional valence and the evaluation value using the evaluation term "pleasant," as specified in step S21.

In this way, it is possible to estimate property information that exhibits the desired relational information. This allows for the estimation of the properties of an object that show a specified relationship between specific state information and evaluation information. For example, when designing a product where the evaluation of comfort improves as the evaluator becomes more comfortable, it becomes possible to estimate the properties of such a product. By developing the product with the estimated properties as the target, it is possible to develop a product that exhibits the expected change in evaluation value according to the evaluator's state.

The present invention will be explained in more detail below, citing an example that analyzes the correlation between evaluation values and evaluation conditions using evaluation terms as indicators, and the relationship with the nature of the object being evaluated. It goes without saying that the present invention is not limited to this example.

<Examples>
[1] Measurement of the properties of the objects This example aims to examine the sense of touch as one example of the human senses. In order to examine the sensations of a wide range of substances that are touched in daily life, a finger-shaped measuring device (Syntouch's Toccare system) capable of measuring pressure, temperature, and vibration was used to measure the following 15 types of physical properties for five types of objects: cashmere, abrasive sponge, artificial skin, beauty serum, and cream. These physical properties were selected mainly to affect the sense of touch.

The 15 physical properties measured are as follows: fST: Tactile Stiction, fRS: Sliding Resistance, mTX: Macrotexture, mCO: Macrotexture Coarseness, mRG: Macrotexture Regularity, μRO: Microtexture Roughness, μCO: Microtexture Coarseness, aTK: Adhesive Tack, tCO: Thermal Cooling, tPR: Thermal Persistence, cCM: Tactile Compliance, cDF: Local Deformation, cDP: Damping, cRX: Relaxation, cYD: Yielding.

[2] Selection of Evaluation Terms In this embodiment, evaluators who touched the object were first asked to freely describe the feel and the feelings and emotions they experienced at that time, and words that were thought to have been obtained through touch were extracted from these responses. Furthermore, from the extracted words, terms that were reported in a previous study (Keitaro Kuramitsu, Journal of the Japan Society of Cosmetic Scientists, Vol. 49 (2015), No. 4, pp. 319-327) that describes the physical characteristics of a substance and creates a glossary of terms for evaluating touch on the same evaluation axis were classified as evaluation terms representing lower-order sensitivities. In addition, terms other than evaluation terms representing lower-order sensitivities that are related to impressions and emotion recall were classified as evaluation terms representing higher-order sensitivities.

From the selected evaluation terms, we further narrowed it down to 22 terms. For the lower-order evaluation terms, we selected those directly related to physical properties, while for the higher-order evaluation terms, we selected those more closely related to emotional valence (representing pleasure/displeasure) and arousal levels. Specifically, the following evaluation terms were selected:

Ten words—"soft," "hard," "smooth," "rough," "silky," "uneven," "cold," "warm," "moist," and "sticky"—were defined as evaluation terms representing lower-level sensitivities that directly describe the physical characteristics of an object. Additionally, twelve words—"luxury," "pleasant," "gentle," "warmth," "like," "high quality," "happy," "calming," "comfortable," "irritating," "disgusting," and "dislike"—were defined as evaluation terms representing higher-level sensitivities that describe impressions evoked by the physical characteristics of an object and lower-level sensitivities. In this example, a glossary of evaluation terms was created based on this lower-level/higher-level classification.

[3] Data acquisition (1) Subjects The subjects were 46 healthy Japanese individuals in their 20s to 50s, with an equal proportion of men and women and different age groups.

(2) Changes in the subject's state due to stimulation In this embodiment, the subject's state was changed by applying stimuli, and the same subject was repeatedly asked to provide evaluation values for each object using the same evaluation terms as indicators. Specifically, five types of images were selected from the emotion-evoking stimulus image database OASIS, representing five quadrants: anger, happiness, normalcy, relaxation, and sadness, and these were presented to the subject using a display before they touched the object to be evaluated.

(3) Acquisition of emotional valence and arousal level After the presentation of stimuli, participants were asked to use an Affective Slider to evaluate which of the following conditions was closer to their current state: emotional valence, which represents pleasure or displeasure, and arousal level, which represents calmness or alertness. This allowed us to acquire the participants' emotional valence and arousal level as conditional information representing the state condition, using integers from 0 to 100 on a scale of 101.

(4) Acquisition of evaluation values For each subject, participants were asked to rate the intensity of 22 evaluation terms for one presented object using the Visual Analogue Scale (VAS) method. Prior to the test, participants practiced answering using the VAS method and were instructed not to make judgments based on factors other than touch, such as sight or smell. Intensity, as used here, is a measure of how strongly the impression or sensation represented by the evaluation term is felt in relation to the object.

In this experiment, participants were instructed to touch all objects placed in a black box, preventing them from seeing the objects. Furthermore, touching was limited to the pad of the index finger of the non-dominant hand, and participants were required to always be touching the object while evaluating the terms. Participants answered the terms by operating a mouse with their dominant hand, responding to terms displayed on the PC monitor. They wiped their fingers between each object exchange to ensure consistent finger condition at the start. Prior to the test, participants were instructed to practice all operations and to avoid making judgments based on factors other than touch, such as sight or smell.

[4] Correlation Analysis A correlation analysis was conducted for each subject between the emotional value, expressed by a combination of emotional valence and arousal level, and the evaluation value for each evaluation term. For each subject, the correlation coefficient and p-value were calculated for the correlation between the psychological score and the evaluation value for each object.

Furthermore, for each object, the results of multiple subjects were categorized as either positive or negative based on the correlation coefficient, and the mean correlation coefficient for each subject was calculated. Additionally, subjects with an absolute value of 0.4 or higher for the correlation coefficient between psychological scores and evaluation values were considered to have a correlation, and the number of subjects with a correlation was identified for each object.

Furthermore, correlation analysis was performed on each of the correlation coefficients and the number of subjects with correlation, based on the physical properties of the objects. The results of the correlation analysis between each of the 15 physical properties listed above, the correlation coefficient, and the number of subjects with correlation for each object showed significant correlations between the physical properties and the psychological scores and evaluation values in the combinations shown in Figures 6 and 7.

Figure 6 shows the average correlation coefficients between emotional valences (representing pleasure/displeasure) and evaluation values for the "pleasant" evaluation term, and the combinations of each physical property value that showed a correlation of a certain level or higher. The average correlation coefficient represents the average of the positive and negative values that showed correlation.

Figure 7 also shows the number of subjects who showed a correlation between emotional valence (representing pleasure/displeasure) and evaluation values for various evaluation terms, and the combinations of each physical property that showed a correlation of a certain level or higher.

[5] Conclusion From the above, it became clear that there are combinations in which a correlation is observed between information on the correlation between evaluation values and evaluation conditions (especially the mental state of the subjects) using evaluation terms as indicators, and the physical properties of the objects being evaluated. Furthermore, it was suggested that the relationship between the correlation between evaluation values and evaluation conditions and the properties of the objects being evaluated can be analyzed using a similar method.

1: Estimation device 11: Property information acquisition means 12: Property relationship information acquisition means 13: Estimation means 2: Estimation device 21: Relationship information specification means 22: Property relationship information acquisition means 23: Estimation means

Claims (18)

An estimation method for estimating information regarding changes in subjective evaluation of a subject according to the state of the evaluator, based on the properties of the subject,
A step of acquiring relationship information relating to the relationship between state information relating to the state of the evaluator and evaluation information indicating the evaluator's subjective evaluation of the subject, and property relationship information indicating the relationship between the property information indicating the property of the subject,
The steps include obtaining property information about the object to be estimated,
An estimation method in which a computer performs the steps of: estimating the relational information from the property information of the object to be estimated, based on the aforementioned property relational information. An estimation method for estimating the properties of an object such that the subjective evaluation of the object shows a specific change depending on the state of the evaluator,
A step of acquiring relationship information relating to the relationship between state information relating to the state of the evaluator and evaluation information indicating the evaluator's subjective evaluation of the subject, and property relationship information indicating the relationship between the property information indicating the property of the subject,
A step of specifying relationship information that indicates the desired relationship,
An estimation method in which a computer performs the steps of estimating the property information of an object that corresponds to the specified relationship information based on the property relationship information. The estimation method according to claim 1 or claim 2, wherein the aforementioned property information is a value representing the properties of the object. The estimation method according to claim 3, wherein the subject is an object, and the property information is the physical property value of the object. The estimation method according to claim 4, wherein the aforementioned physical property value is a value that represents a property perceptible to human touch. The estimation method according to any one of claims 1 to 5, wherein the state information includes a state value indicating the mental state of the evaluator at the time of evaluation. The estimation method according to claim 6, wherein the state value includes an emotional value relating to the evaluator's emotions. The estimation method according to claim 7, wherein the aforementioned emotional value includes an emotional valence indicating the degree of pleasure or displeasure of the evaluator. The estimation method according to claim 7, wherein the emotion value includes the evaluator's level of arousal. The estimation method according to claim 7, wherein the aforementioned emotional value is represented by a combination of the emotional valence, which indicates the degree of pleasure or displeasure of the evaluator, and the evaluator's level of arousal. The estimation method according to any one of claims 1 to 10, wherein the relational information includes a value relating to the magnitude of the correlation between the evaluation information and the state information. The estimation method according to any one of claims 1 to 11, wherein the relational information includes a value relating to the number or percentage of evaluators showing a correlation between the evaluation information and the status information. The estimation method according to any one of claims 1 to 12, wherein the evaluation information is an evaluation value for the subject based on the evaluator's subjective assessment, using evaluation terms as indicators. The estimation method according to claim 13, wherein the evaluation value is an evaluation value of sensitivity indicating the strength with which the evaluator perceives the sensation represented by the evaluation term. An estimation device for estimating information regarding changes in subjective evaluation of an object in accordance with the evaluator's state, based on the properties of the object,
Means for acquiring property relationship information, which indicates the relationship between status information relating to the state of the evaluator and evaluation information indicating the evaluator's subjective evaluation of the subject, and property information indicating the properties of the subject.
Regarding the object to be estimated, a means for obtaining property information,
An estimation device comprising means for estimating the relational information from the property information of the object to be estimated, based on the aforementioned property relational information. An estimation device for estimating the properties of an object such that the subjective evaluation of the object shows a specific change depending on the state of the evaluator,
Means for acquiring property relationship information, which indicates the relationship between status information relating to the state of the evaluator and evaluation information indicating the evaluator's subjective evaluation of the subject, and property information indicating the properties of the subject.
Means for specifying relational information that indicates a desired relationship,
An estimation device comprising means for estimating the property information of an object corresponding to the specified relationship information based on the property relationship information. An estimation program for estimating information regarding changes in subjective evaluation of an object according to the evaluator's state, based on the properties of the object, wherein the computer...
A step of acquiring relationship information relating to the relationship between state information relating to the state of the evaluator and evaluation information indicating the evaluator's subjective evaluation of the subject, and property relationship information indicating the relationship between the property information indicating the property of the subject,
The steps include obtaining property information about the object to be estimated,
An estimation program that performs the steps of estimating the relational information from the property information of the object to be estimated, based on the aforementioned property relational information. An estimation program for estimating the properties of an object such that the subjective evaluation of the object shows a specific change depending on the state of the evaluator, wherein the computer...
A step of acquiring relationship information relating to the relationship between state information relating to the state of the evaluator and evaluation information indicating the evaluator's subjective evaluation of the subject, and property relationship information indicating the relationship between the property information indicating the property of the subject,
A step of specifying relationship information that indicates the desired relationship,
An estimation program that performs the steps of: estimating the property information of an object corresponding to the specified relationship information based on the property relationship information.