JP2022047253A - Impression estimation device, impression estimation method, and impression estimation program - Google Patents

Abstract

To improve estimation accuracy of own impression from others by reflecting the psychological state that cannot be properly evaluated by estimation methods that employ appearance as an indicator.SOLUTION: An impression estimation device is provided for estimating own impression from others. The impression estimation device comprises: a feature quantity acquisition unit for acquiring a feature quantity of own face; an evaluation value acquisition unit for acquiring an evaluation value of self-impression; and an estimation unit for estimating impression using an impression estimation formula prepared in advance to estimate impression from others using a feature quantity and an evaluation value of self-impression as variables, and the acquired feature quantity and evaluation value.SELECTED DRAWING: Figure 3

Description

The present disclosure relates to techniques for estimating one's own impression by others.

What kind of impression you give to others is one of the important factors in communicating with others. Conventionally, for example, an evaluation method for quantifying one's personality and psychological state by using a question sheet has been known. Further, there is known a technique of estimating the impression of a subject held by another person by using the appearance of the subject as an index (for example, Patent Document 1).

Japanese Unexamined Patent Publication No. 2015-527224

However, the impression that oneself gives to others is influenced not only by the appearance but also by one's inner side, that is, the atmosphere created by the psychological state. With the conventional estimation technology using the appearance as an index, it is not possible to evaluate changes in the appearance that are affected by the psychological state, and depending on the evaluation method using the appearance as an index, the impression that oneself gives to others can be estimated accurately. Couldn't.

Therefore, it is desired to improve the estimation accuracy of one's own impression by others by reflecting the psychological state that cannot be appropriately evaluated by the estimation method using the appearance as an index.

In order to solve at least a part of the above problems, the present disclosure adopts the following various aspects.

The first aspect provides an impression estimation device that estimates one's own impression by another person. The impression estimation device according to the first aspect has a feature amount acquisition unit that acquires a feature amount of its own face, an evaluation value acquisition unit that acquires an evaluation value of its own self-impression, and an evaluation value of the feature amount and its own impression. It is provided with an impression estimation formula prepared in advance for estimating an impression by another person, and an estimation unit for estimating an impression using the acquired feature amount and the evaluation value.

According to the impression estimation device according to the first aspect, the feature amount and the evaluation value of the self-impression are used as variables, and the impression estimation formula prepared in advance for estimating the impression by another person and the acquired feature amount are used. Since the impression is estimated using the evaluation value, it is possible to improve the estimation accuracy of one's own impression by another person by reflecting the psychological state that cannot be appropriately evaluated by the estimation method using the appearance as an index.

In the impression estimation device according to the first aspect, the impression estimation formula is a formula calculated by regression analysis with the feature amount and the evaluation value of self-impression as explanatory variables and the impression value of another person as an objective variable. There may be. In this case, it is possible to estimate an impression that reflects the feature amount, the self-impression, and the impression of another person.

In the impression estimation device according to the first aspect, the impression value by the other person may be determined by the evaluation of the degree of the self-impression obtained from the own face. In this case, it is possible to obtain the impression value of another person based on his / her appearance.

In the impression estimation device according to the first aspect, the evaluation value acquisition unit may acquire a factor score obtained by selecting a plurality of factors associated with the impression as the evaluation value. In this case, the evaluation value of self-impression can be obtained more objectively.

In the impression estimation device according to the first aspect, the evaluation value acquisition unit identifies the factor by a term indicating a psychological state associated with the impression selected by itself, and the factor loading of the identified factor. The factor score calculated by using the quantity may be acquired as the evaluation value. In this case, since the term indicating the psychological state is used, the evaluation value of the self-impression can be obtained easily and more objectively.

In the impression estimation device according to the first aspect, the evaluation value acquisition unit determines the degree of self-evaluation for terms selected by itself and associated with a plurality of factors indicating a psychological state associated with the impression. The factor score calculated by using the factor loading amount associated with the term may be acquired as the evaluation value. In this case, the evaluation value can be easily calculated.

In the impression estimation device according to the first aspect, even if the feature amount acquisition unit acquires at least the feature amount related to the eyes, the feature amount related to the nose, and the feature amount related to the mouth of its own face as the feature amount of the face. good. In this case, it is possible to improve the acquisition accuracy of the feature amount of the self-appearance.

The second aspect provides an impression estimation method for estimating one's own impression by another person. In the impression estimation method according to the second aspect, the feature amount of one's own face is acquired, the evaluation value of self-impression by oneself is acquired, and the impression by another person with the feature amount and the evaluation value of self-impression as variables is obtained. It includes an impression estimation formula prepared in advance for estimation, and an impression estimation using the acquired feature amount and the evaluation value.

According to the impression estimation method according to the second aspect, the feature amount and the evaluation value of the self-impression are used as variables, and the impression estimation formula prepared in advance for estimating the impression by another person and the acquired feature amount are used. Since the impression is estimated using the evaluation value, it is possible to improve the estimation accuracy of one's own impression by another person by reflecting the psychological state that cannot be appropriately evaluated by the estimation method using the appearance as an index.

A third aspect provides an impression estimation program that estimates one's own impression by others. The impression estimation program according to the third aspect has a function for acquiring the feature amount of one's own face, a function for acquiring an evaluation value of self-impression by oneself, and an evaluation value of the feature amount and self-impression. A computer realizes an impression estimation formula prepared in advance for estimating an impression by another person as a variable, and a function for estimating an impression using the acquired feature amount and the evaluation value.

According to the impression estimation program according to the third aspect, the feature amount and the evaluation value of the self-impression are used as variables, and the impression estimation formula prepared in advance for estimating the impression by another person and the acquired feature amount are used. Since the impression is estimated using the evaluation value, it is possible to improve the estimation accuracy of one's own impression by another person by reflecting the psychological state that cannot be appropriately evaluated by the estimation method using the appearance as an index.

The impression estimation program according to the third aspect may be stored in a computer-readable medium such as a CD, DVD, Blu-ray disc, or semiconductor memory.

Explanatory drawing which shows an example of the impression estimation apparatus and the impression estimation system which concerns on 1st Embodiment. Explanatory drawing which shows an example of the impression estimation apparatus and the impression estimation system which concerns on 1st Embodiment. Explanatory drawing which shows the impression estimation system which concerns on 1st Embodiment by a functional part. The block diagram which shows the functional internal structure of the impression estimation apparatus which concerns on 1st Embodiment. An explanatory diagram showing an example of factors, terms, and factor loadings associated with an impression. An explanatory diagram showing an example of terms collected as terms associated with an impression. An explanatory diagram showing an example of the main factors associated with an impression and the terms associated with each factor. An explanatory diagram showing an example of a questionnaire sheet used to determine the main factors associated with an impression and to acquire one's psychological state. A graph showing the correlation between the estimated impression value calculated using the feature amount and the evaluation value of the self-impression and the sensory other person's impression value. A graph showing the correlation between the estimated impression value calculated using the feature amount and the sensory other person's impression value. A graph showing the correlation between the estimated impression value calculated using the evaluation value of self-impression and the impression value of a sensory stranger. The flowchart which shows each processing step of the impression estimation processing executed by the impression estimation apparatus which concerns on 1st Embodiment.

In the conventional impression estimation by others, the impression estimation by others has been performed based on the evaluation of oneself, that is, the photographed image of the subject. In the captured image, it is not possible to express not only the appearance but also the inner surface of oneself, that is, the atmosphere created by the psychological state, which can be felt by others when face-to-face. Or, it could not be estimated accurately. The present inventors have diligently examined the reflection of the psychological state of the subject, which is not properly expressed in the captured image of the subject, in the impression estimation by others. Hereinafter, the impression estimation device, the impression estimation method, and the impression estimation program according to the present disclosure will be described with reference to the drawings, based on the embodiments.

First embodiment:
1 and 2 show an example of an impression estimation device and an impression estimation system according to the first embodiment. The impression estimation device 100 may be realized as a tablet terminal or a smartphone (hereinafter, simply referred to as “smartphone”) as shown in FIG. 1, or may be realized as a desktop computer as shown in FIG. Further, it may be realized as a notebook computer (not shown). When the smartphone 100 functions as an impression estimation device, it includes an image pickup unit 20 and a display unit 21, and constitutes an impression estimation system 10 by itself. The display unit 21 can also function as an input device for inputting text and selection. In the smartphone 100, the image pickup unit 20 captures a face image or an upper body image of the target person whose impression is estimated by another person, that is, the operator's own body image, and the display unit 21 displays the impression estimation result. Impression estimation results include, for example, the impression that they look shining, have a sense of responsibility, and are natural. A full-body image may be captured instead of the face image or the upper body image, but in order to obtain the facial expression or feature extraction accuracy suitable for estimating the impression, the face image or the upper body image, and further It is desirable that a facial image be imaged. On the other hand, when the computer 100 functions as an impression estimation device, the image pickup device 30 and the display device 31 are separately used, and the impression estimation system 10 is configured by the computer terminal 100, the image pickup device 30 and the display device 31. A keyboard 32 for inputting text and a mouse 33 for inputting selection are connected to the computer terminal 100. The image pickup apparatus 30 captures a face image or an upper body image of itself to be evaluated. As the image pickup apparatus 30, a general digital camera or a smartphone may be used. The display device 31 displays the impression estimation result. The computer 100 and the image pickup device 30 are connected by wire or wirelessly, and image pickup data can be transmitted from the image pickup device 30 to the computer 100. Further, since the tablet terminal or smartphone can be used as a display device, the impression estimation result may be transmitted from the computer 100 to the tablet terminal or smartphone by wireless communication. In this case, the impression estimation result can be transmitted to the user's smartphone. The wireless connection is realized by wireless communication conforming to a short-range wireless communication standard such as Wifi (registered trademark) or Bluetooth (registered trademark). The computer 100 may be a server computer that functions as a server connected to a plurality of terminals via a wired or wireless network. In this case, skin images to be evaluated from a plurality of various client terminals. The data is uploaded, the impression estimation value is calculated in the server computer, and the calculated impression estimation value can be transmitted from the server computer to various client terminals.

FIG. 3 is an explanatory diagram showing the impression estimation system 10 according to the first embodiment by a functional unit. The impression estimation device 100 includes a feature amount acquisition unit F1, an evaluation value acquisition unit F2, and an estimation unit F3. The feature amount acquisition unit F1 acquires face image data from the image pickup unit 20 or the image pickup device 30, and acquires the feature amount from the acquired face image data. The evaluation value acquisition unit F2 calculates a factor score using a plurality of factors associated with the impression, which are input from the input devices 21, 32, and 33 by the subject, and acquires or associates the factor score as the evaluation value. Factors are specified using terms that indicate the psychological state of the person, and the factor score is calculated and obtained as an evaluation value. The evaluation value acquisition unit F2 further determines the degree of self-evaluation for terms input by the subject from the input devices 21, 32, 33 and associated with a plurality of factors indicating the psychological state associated with the impression. The factor score can be calculated using the factor loading associated with the term and obtained as an evaluation value. The estimation unit F3 estimates the impression using the feature amount acquired by the feature amount acquisition unit F1 and the evaluation value acquired by the evaluation value acquisition unit F2, and an impression estimation formula using the feature amount and the evaluation value as variables. The estimated impression estimated by the estimation unit F3 is output to the display unit 21 or the display device 31.

FIG. 4 is a block diagram showing a functional internal configuration of the impression estimation device 100 according to the first embodiment. The impression estimation device 100 includes a central processing unit (CPU) 101, a memory 102, an input / output interface 103, and a clock generator (not shown). The CPU 101, the memory 102, the input / output interface 103, and the clock generator are connected so as to be communicable in both directions via the internal bus 104. The memory 102 includes a memory for storing the impression estimation program Pr1 for estimating one's own impression by another person in a non-volatile manner and read-only, for example, a ROM and a memory for reading and writing by the CPU 101, for example, RAM. The non-volatile and readable / writable area of the memory 102 includes the evaluation term storage area 102a for storing the evaluation term for obtaining the evaluation value. The evaluation term is, for example, a term indicating a psychological state associated with an impression, and each term is associated with a plurality of factors that define the impression. More specifically, for example, FIG. 5 shows a plurality of factors 1 to 6 associated with the impression that others have toward themselves, and the psychology associated with each factor. An example of the correspondence with the term indicating the state is shown. Each term is assigned a factor loading that indicates the intensity that affects each factor. The factor contribution rate shown at the bottom of each factor indicates the ratio of each factor to the impression "looks shining". Each factor 1 to 6 and the factor loading are determined by the procedure described later.

The CPU 101, that is, the impression estimation device 100 expands the impression estimation program Pr1 stored in the memory 102 into a readable / writable memory and executes the feature amount acquisition unit F1, the evaluation value acquisition unit F2, and the estimation unit F3. Functions as. The CPU 101 may be a single CPU, a plurality of CPUs that execute each program, or a multitasking type or multithread type CPU capable of simultaneously executing a plurality of programs. Is also good. The feature amount acquisition unit F1, the evaluation value acquisition unit F2, and the estimation unit F3 may be realized as a logic circuit in terms of hardware.

The input / output interface 103 is a physical and logical interface used for connecting the impression estimation device 100 and various other devices, and is, for example, a known terminal such as a USB terminal, a LAN terminal, or a serial bus terminal. It can be realized. In the first embodiment, the input / output interface 103 is connected to, for example, a display unit 21 having a liquid crystal panel, a display device 31, for example, an image pickup unit 20 having an image pickup element, and an image pickup device 30 via signal lines. .. In addition, instead of the display device 31, or together with the display device 31, a printing device that outputs, that is, prints, an estimated impression on a printing medium, for example, paper may be provided.

Determining the impression estimation formula:
The procedure for determining the terms indicating the psychological state associated with the "look shining" impression, the factors 1 to 6 described above, and the factor loading amount will be described. First, for 124 women, we collected terms or words that indicate "a state in which oneself is shining" and terms or words that indicate "a state in which others appear to shine", and used text mining software such as KH Coder. Then, a term indicating a psychological state, which is a frequent word of 66 words shown in FIG. 6, that is, a rating term was obtained. Note that each term shown in FIG. 6 is an example, and some terms are omitted in FIG. Next, 109 women were asked to evaluate each of the 66 rating terms on a 6-point scale as to whether or not they fit into the "state in which they are shining" and the "state in which others appear to be shining." .. The six stages are very well applicable, applicable, rather applicable, rather not applicable, not applicable, and not applicable at all. Factor analysis was performed on the obtained answers using 66 rating terms as observation variables, and 6 factors associated with the impression of "looking shining" were determined. The six factors are factor 1 (peace of mind), factor 2 (enthusiasm for work), factor 3 (facial expression honesty), factor 4 (consciousness to see), and factor 5 (satisfaction with oneself) shown in FIG. Feeling), factor 6 (awareness of contribution to others). Further, as shown in FIG. 7, three terms having a high factor loading amount were extracted from the terms indicating the psychological state associated with each factor.

As shown in FIG. 8, 17 women were targeted, and as shown in FIG. 8, the three terms associated with each factor, that is, the 18 words indicating the psychological state and the degree to each term, as described above, were 6 levels. I asked them to answer their current psychological state using a questionnaire sheet containing. Of the degrees obtained from the answers, the factor loadings shown in FIG. 5 were applied and added to the terms indicating the psychological states that were considered to be very well applicable, and the factor scores were obtained. That is, a term that fits very well was set to 1, other terms were set to 0, and the factors were multiplied and added to obtain a factor score. It should be noted that each degree is weighted, for example, multiplied by (5 points) that is very well applied (5 points) but not (0 points) at all, and the normalized (1/5) value is multiplied by the factor loading and added. Is also good. In addition, along with the answers to the questionnaire sheet, the facial photograph of the subject was taken and the feature amount of the facial image was extracted. Specifically, the average value of the area of the right eye and the area of the left eye, the ratio occupied by the pupil when the area of the entire eye is 1, (the area ratio of the pupil), the center of the boundary from the base of the nose to the center of the upper lip. The length of the tie (the length under the nose), the area connecting the fissure and the corner of the mouth, and the area of the lips are extracted as feature quantities. In addition, at least a part of these feature amounts and other feature amounts may be additionally or used instead.

Finally, an impression estimation formula using the evaluation value of self-impression and the feature amount of the facial image as explanatory variables and the impression value of others by others as the objective variable is calculated by regression analysis. Five evaluators were asked to perform an impression evaluation as to whether or not the face image acquired above "looks shining". Each evaluator gave 1 point for each facial image if it had the impression that it looked shining, -1 point if it had the impression that it did not look shining, and if it did not have either impression. 0 points are assigned to. The average value of the points assigned by the five evaluators was calculated and used as the impression value of the other person. A partial minimum squared regression (PLS) analysis was performed using the factor scores and features of the six factors obtained for each face image as explanatory variables and the sensory other person's impression value as the objective variable, and the factors of the six factors. The regression coefficient and the constant term were obtained from the score and the feature quantity. More specifically, using the leave-one-out cross-validation method, an impression estimation formula is calculated and calculated using 16 samples out of 17 samples (features, self-impression evaluation values, and impression values of others). The number of components (dimensions) suitable for PLS analysis using the estimated impression value obtained by using the impression estimation formula, the feature amount of the remaining 1 sample, and the evaluation value of self-impression, and the impression value of others in the remaining 1 sample. Number) is determined. Based on the determined number of components, PLS analysis is performed using 17 samples of features and self-impression evaluation values, as well as the corresponding 17 samples of others' impression values, to determine the regression coefficient, i.e., the factor loading. Then, the following impression estimation formula (Equation 1) for estimating one's own impression by another person was obtained. It should be noted that each feature amount is a factor score for calculating an impression estimation formula, and the impression estimation formula may be obtained by another regression method that approximates a straight line.
Estimated impression value = 0.030582835 x "peace of mind" factor score + 0.01365280 x "enthusiasm for work" factor score + 0.03269904 x "honestness of expression" factor score + 0.022219861 x "awareness of appearance" factor score + (-0.00194219) x "Satisfaction with oneself" factor score + 0.041448159 x "Contribution awareness to others" factor score + 0.00000971 x Area of the entire eye (pixel) + (-0.00035344) x Eyes Area ratio (%) + (-0.00179344) x length under the nose (pixel) + 0.00000447 x mouth cleft angle area (pixel) + (-0.00000127) x lip area (pixel) + (- 1.58179073) (constant term) ... (Equation 1)

Using 17 samples, the correlation between the estimated impression value obtained from the impression estimation formula shown in Equation 1 and the above-mentioned sensory other person's impression value was verified. FIG. 9 is a graph showing the correlation between the evaluation value of self-impression, that is, the estimated impression value and the sensory other person's impression value when the factor score and the feature amount are used. The correlation coefficient between the estimated impression value and the sensory other person's impression value in this case was R = 0.678.

For comparison, the impression estimation formula obtained by the regression analysis when only the feature amount is used and when only the factor score which is the evaluation value of the self-impression is used will be verified below. The impression estimation formula (2) when only the feature amount is used is as follows.
Estimated impression value = 0.00002102 x area of the entire eye (pixel) + (-0.00076561) x area ratio of the pupil (%) + (-0.00388486) x length under the nose (pixel) + 0.00000969 x Mouth fissure angle area (pixel) + (-0.00000276) x lip area (pixel) + 0.094886807 (constant term) ... (Equation 2)
Using 17 samples, the correlation between the estimated impression value obtained from the impression estimation formula shown in Equation 2 and the above-mentioned sensory other person's impression value was verified. FIG. 10 is a graph showing the correlation between the estimated impression value and the sensory other person's impression value when the feature amount is used. The correlation coefficient between the estimated impression value and the sensory other person's impression value in this case was R = 0.657.

The impression estimation formula (3) when only the factor score is used is as follows.
Estimated impression value = 0.032368826 x "Peace of mind" factor score + 0.01445008 x "Enthusiasm for work" factor score + 0.03460856 x "Facial honesty" factor score + 0.02344944 x "Appearance awareness" factor Score + (-0.00205561) x "Satisfaction with oneself" factor score + 0.0434398 x "Contribution awareness to others" factor score + (-1.71419546) (constant term) ... (Equation 3)
Using 17 samples, the correlation between the estimated impression value obtained from the impression estimation formula shown in Equation 3 and the above-mentioned sensory other person's impression value was verified. FIG. 11 is a graph showing the evaluation value of self-impression, that is, the correlation between the estimated impression value and the sensory other person's impression value when the factor score is used. The correlation coefficient between the estimated impression value and the sensory other person's impression value in this case was R = 0.525.

As described above, it was confirmed that the correlation coefficient R was increased by using the evaluation value of self-impression in addition to the feature amount. Therefore, by estimating one's own impression by another person using the evaluation value of one's own impression in addition to the feature amount, one's psychological state is reflected and the estimation accuracy of one's own impression by another person can be improved. It was verified.

Estimating one's own impression by others using an impression estimator:
The impression estimation process executed by the impression estimation device 100 will be described. FIG. 12 is a flowchart showing each processing step of the impression estimation process executed by the impression estimation device 100 according to the first embodiment. More specifically, the CPU 101 is realized by executing the impression estimation program Pr1.

The CPU 101 acquires face image data of itself (hereinafter, also referred to as “target person”) who wants to know the impression of another person via the input / output interface 103 (step S100). The face image data is acquired by the subject using the image pickup device 30 or the image pickup unit 20. In addition, instead of the face image data, the upper body image data of the subject may be acquired.

The CPU 101 acquires a feature amount from the acquired face image data (step S102). To acquire the feature amount, first, the feature extraction from the face image data is a known method, for example, the parts that characterize the face such as eyes, nose, and mouth by a method such as edge extraction based on the brightness difference or region extraction using the color value. Is extracted, and then layout information such as the position and area of each part is extracted. In this embodiment, the average value of the area of the right eye and the area of the left eye, the ratio occupied by the pupil when the area of the entire eye is 1, (the area ratio of the pupil), the center of the boundary from the base of the nose to the center of the upper lip. The length connecting the two (the length under the nose), the area connecting the fissure and the corner of the mouth, and the area of the lip are obtained as feature quantities. When other features are added or used instead in the impression estimation formula, the corresponding other features are acquired.

The CPU 101 acquires an evaluation value indicating the psychological state of the subject (step S104). The evaluation value indicating the psychological state of the subject, that is, the evaluation value of self-impression by oneself, is the evaluation value obtained by calculating the factor score using a plurality of factors associated with the impression, or associated with the impression. Factors are identified using terms that indicate the psychological state of the person, the evaluation value obtained by calculating the factor score, and self-evaluation of the terms associated with multiple factors that indicate the psychological state associated with the impression. It includes the evaluation value obtained by calculating the factor score using the degree and the factor loading associated with each term. The evaluation value of the self-impression by oneself in the present embodiment can be said to be the factor score of a plurality of factors indicating the psychological state associated with the impression obtained by the self-evaluation by oneself. The acquisition of the evaluation value is performed, for example, by displaying the questionnaire sheet shown in FIG. 8 on the display unit 21 or the display device 31, and selecting by the target person, that is, inputting an answer. The questionnaire sheet contains a plurality of terms indicating the target impression, and in this embodiment, the psychological state associated with "looks shining", and the degree of the target person at the present time for each term, that is, By inputting the degree of feeling, the factor scores of the six factors shown in FIG. 7 can be obtained as evaluation values. Specifically, by multiplying the degree of 0 to 5 for the three terms included in each factor by the factor loading of each factor shown in FIG. 5 as a coefficient and normalizing (1/5), each factor is used. The factor score of the factor, that is, the evaluation value is obtained. In addition to this, the acquisition of evaluation values includes the six factors shown in Fig. 7, "I feel at ease," "I am enthusiastic about my work," "I am conscious of what I see," and "To myself." It may be displayed as a factor such as "satisfied" or "helping others", and may be executed by selecting a factor that applies to the subject or selecting the degree to which the subject applies. In this case, the factor contribution rate of each factor shown in FIG. 5 is used as a coefficient, and the factor score of each factor is evaluated by multiplying the value of 0 or 1 obtained as a result of selection or the numerical value of 0 to 5. Obtained as a value. Further, the subject may select a plurality of terms corresponding to the 6 factors, and the corresponding factor may be selected from the 6 factors according to the selected term.

The CPU 101 acquires an estimated impression value using the above-mentioned impression estimation formula prepared in advance (step S106). Specifically, the estimated impression value is calculated using the feature amount acquired in steps S102 and S104 and the factor score of each factor which is the evaluation value. The calculated estimated impression value takes a value between -1 and +1. +1 means that the impression of "looking shining" is the strongest, and "-1" means that the impression of "looking shining" is the strongest. It means weak.

The CPU 101 displays the acquired estimated impression value on the display unit 21 or the display device 31 (step S108), and ends the processing routine. The estimated impression value displayed on the display unit 21 or the display device 31 may be replaced with a value of 0 to 100 or a value of 0 to 10, or -1 to +1 may be displayed in 5 stages. It may be divided and displayed by being replaced with words such as "getting the strongest impression", "getting the impression", "neither", "not getting the impression", and "getting the weakest impression".

As described above, according to the impression estimation device 100 according to the present embodiment, when estimating one's own impression by another person, in addition to the feature amount obtained from one's own face image, an evaluation value of one's own impression is obtained. Since the impression estimation formula as a variable is used, it is possible to improve the estimation accuracy of one's own impression by another person by reflecting the psychological state that cannot be appropriately evaluated by the estimation method using the appearance as an index. More specifically, in the first embodiment, when estimating whether or not another person has the impression of "looking shining", it is associated with the impression of "looking shining". An impression estimation formula is used in which one's own psychological state is evaluated using terms indicating the psychological state, and the feature amount of the facial image and a plurality of factors indicating the psychological state are used as explanatory variables. As a result, it is possible to estimate whether or not others have the impression that they "look shining" by reflecting their own psychological state that cannot be obtained from the features of their own facial images. Become. That is, although it is difficult to obtain by general feature extraction using a general image pickup device, the psychological state appearing in the appearance, which is obtained when another person faces himself / herself, is reflected in the estimation of the impression. Therefore, the discrepancy between the impression that another person has when facing oneself and the estimated impression when oneself estimates the impression by another person is reduced or eliminated.

In the impression estimation device 100 according to the first embodiment, the evaluation value of one's own psychological state selects a factor corresponding to the psychological state associated with the impression to be estimated, or determines the degree of each factor. In this case, the subject can input his / her psychological state through simple choices, so that he / she can easily estimate his / her impression by others. On the other hand, when the evaluation value of one's own psychological state is acquired by using multiple terms associated with the psychological state, it becomes possible to acquire one's own psychological state more objectively, and one's own by another person. Impression estimation accuracy is improved.

According to the impression estimation device 100 according to the first embodiment, the accuracy of estimating one's own impression by another person can be improved, so that a makeup method, a massage method, or a stretch for eliminating the deviation from the recognition by oneself can be improved. The law can be proposed to the subject. For example, if the score of "peace of mind" is low among the factors of the evaluation value of one's psychological state, relaxing skin care, fragrance, and dietary suggestions are offered, and the score of "enthusiasm for work" is low. Suggestions for fragrances that enhance feelings, suggestions for stretching facial muscles when the score for "facial expression is straightforward", and stretches for improving posture when the score for "awareness of appearance" is low. If the score of "Skin care" or "Contribution to others" is low, you can make a calming scent.

Other embodiments:
(1) In the above embodiment, the obtained impression estimation value may be indicated by a numerical value or may be indicated by a graph. When displayed as a graph, it is possible to urge the user to intuitively grasp the impression estimation, and when displayed numerically, it is possible to show the impression estimation more concretely and directly. can.

(2) In the above embodiment, "looks shining" is used as the estimated impression, but other impressions such as "gentle", "bright", "refreshing", and "healthy" are used. May be done. In this case, it is possible to obtain the evaluation value of self-impression by acquiring the terms indicating the psychological state associated with each impression, obtaining the main factors, and creating a questionnaire sheet by the method described above. Will be. After the evaluation value of the self-impression is obtained, the impression estimation formula is calculated using the feature amount of the facial image.

Although the present disclosure has been described above based on various embodiments, the above-described embodiments are for facilitating the understanding of the present disclosure and are not limited to the present disclosure. The present disclosure may be modified or improved without departing from its spirit and claims, and the present disclosure includes its equivalents.

10 ... Impression estimation system, 100 ... Impression estimation device, 101 ... Central processing unit (CPU), 102 ... Storage unit, 20 ... Imaging unit, 21 ... Display unit, 30 ... Imaging device, 31 ... Display device, F1 ... Features Quantity acquisition unit, F2 ... evaluation value acquisition unit, F3 ... estimation unit.

Claims (9)

It is an impression estimation device that estimates one's own impression by others.
The feature amount acquisition unit that acquires the feature amount of one's face,
The evaluation value acquisition unit that acquires the evaluation value of self-impression by oneself,
Estimating the impression using the feature amount and the self-impression evaluation value as variables, the impression estimation formula prepared in advance for estimating the impression by others, and the acquired feature amount and the evaluation value. Department and
An impression estimation device. In the impression estimation device according to claim 1,
The impression estimation formula is an impression estimation device, which is a formula calculated by regression analysis using a feature amount and an evaluation value of self-impression as explanatory variables and another person's impression value as an objective variable. In the impression estimation device according to claim 2,
The impression value of another person is determined by the evaluation of the degree of self-impression obtained from one's own face. In the impression estimation device according to any one of claims 1 to 3,
The evaluation value acquisition unit is an impression estimation device that acquires a factor score obtained by selecting a plurality of factors associated with an impression as the evaluation value. In the impression estimation device according to claim 4,
The evaluation value acquisition unit identifies the factor by a term indicating a psychological state associated with the impression selected by itself, and obtains a factor score calculated by using the factor loading of the identified factor. Impression estimation device to be acquired as an evaluation value. In the impression estimation device according to any one of claims 1 to 3,
The evaluation value acquisition unit determines the degree of self-evaluation for terms associated with a plurality of factors indicating the psychological state associated with the impression selected by itself, and the factor loading amount associated with each term. An impression estimation device that acquires a factor score calculated using the above as the evaluation value. In the impression estimation device according to any one of claims 1 to 6.
The feature amount acquisition unit is an impression estimation device that acquires at least the feature amount related to the eyes, the feature amount related to the nose, and the feature amount related to the mouth of the face as the feature amount of the face. It is an impression estimation method that estimates one's own impression by others.
Get the features of your face and
Obtain the evaluation value of self-impression by yourself,
Estimating an impression using a pre-prepared impression estimation formula that estimates an impression by another person with a feature amount and an evaluation value of self-impression as variables, and the acquired feature amount and the evaluation value.
Impression estimation method. An impression estimation program that estimates one's own impression by others.
With the function to acquire the feature amount of one's face,
A function to acquire the evaluation value of self-impression by oneself,
A function for estimating an impression using a pre-prepared impression estimation formula that estimates an impression by another person with a feature amount and an evaluation value of self-impression as variables, and the acquired feature amount and the evaluation value. An impression estimation program that realizes with a computer.

Images