JP6931840B2 - Light impression evaluation method - Google Patents

Description

The present invention relates to a method for evaluating the impression of light.

Product design is one of the important factors that contribute to the sales of industrial products. Product design exerts functions such as visually motivating users to purchase.

In product design (hereinafter, simply referred to as "design"), in order to give a favorable impression to a user's taste, it is advisable to pay attention to the relationship between the impression given by the design and the user's taste. However, it is difficult to grasp what kind of impression the design gives to the user.

Therefore, conventionally, an impression evaluation method capable of evaluating an impression received by a person has been proposed (for example, Patent Document 1). Patent Document 1 discloses an impression evaluation method capable of evaluating impressions of various objects with high accuracy from physical property information and biological information.

Japanese Unexamined Patent Publication No. 2014-200757

However, since there is diversity among users in the relationship between the impression given by the design and the user's preference, it is difficult to accurately grasp the relationship by the conventional impression evaluation method. In particular, since users have a wide variety of tastes regarding lighting, it is difficult to properly design and plan lighting-related designs. In other words, the light space that each individual feels good about is different, and it is difficult to build a light space that everyone likes.

Therefore, in order to obtain useful knowledge when designing a design, it is conceivable to evaluate the impression of the user by appropriately classifying the user. In this case, users are usually classified using attributes and the degree of preference, but such a classification method cannot properly distinguish users who prefer the same design for different reasons.

The present invention has been made to solve such a problem, and it is possible to classify users based on the relationship after clarifying the relationship between the impression and preference of each user. The purpose is to provide an impression evaluation method for.

In order to achieve the above object, one aspect of the light impression evaluation method according to the present invention includes a step of presenting an image including a light state to a user, a plurality of predetermined impression evaluation words, and a predetermined 1 A user is characterized by a process of calculating a correlation coefficient between a user's preference and an impression on the image using a preference evaluation word representing one preference, and a cluster analysis using the correlation coefficient. Including the step of classifying into clusters by the similarity of the combination of.

According to the present invention, it is possible to clarify the relationship between the impression and preference of each user and then classify the users based on the relationship.

It is a flowchart of the light impression evaluation method which concerns on embodiment. It is a figure which shows the light space image (evaluation target image) used in the light impression evaluation method which concerns on embodiment. It is a figure which shows the impression evaluation word of 9 items used in the impression evaluation method of a light which concerns on embodiment. It is a figure which shows the preference evaluation word used in the light impression evaluation method which concerns on embodiment. It is a figure which shows the evaluation word of 10 items and the option of 7 grade evaluation in the light impression evaluation method which concerns on embodiment. It is a figure which shows an example of the relationship between the correlation coefficient of the user calculated by the light impression evaluation method which concerns on embodiment, and the classified cluster. It is a schematic diagram of an example of the light impression evaluation system for executing the light impression evaluation method which concerns on embodiment.

Hereinafter, embodiments of the present invention will be described. It should be noted that all of the embodiments described below show a preferred specific example of the present invention. Therefore, the numerical values, the components, the arrangement positions and connection forms of the components, the steps (steps), the order of the steps, and the like shown in the following embodiments are examples, and are not intended to limit the present invention. .. Therefore, among the components in the following embodiments, the components not described in the independent claims indicating the highest level concept of the present invention will be described as arbitrary components.

(Embodiment)
The light impression evaluation method according to the embodiment will be described with reference to FIG. FIG. 1 is a flowchart of a light impression evaluation method according to an embodiment.

The light impression evaluation method in the present embodiment is a method for evaluating the impression of the light received by a person from the lighting space, and as shown in FIG. 1, the image presentation step S1, the correlation coefficient calculation step S2, and the cluster classification. Includes step S3. Hereinafter, each step will be described in detail.

[Image presentation process]
In the image presentation step S1, an image including a light state is presented to the user. This image is an evaluation target image used for the user to evaluate the light, and is, for example, a light space image showing a scene of a living space illuminated by the light. As described above, in the present embodiment, as an example in which there is diversity among users in the relationship between the impression given by the design and the user's preference, an experiment of impression evaluation is performed using a light space image in the living space. ..

Specifically, as an image including the state of light (light space image), a photograph showing the state of light in the living space was used. Such a photograph can be obtained by taking a scene of a living space illuminated by a light with a camera or the like. Light The light in the spatial image may be either artificial light or natural light.

In the present embodiment, seven photographs shown in FIGS. 2A to 2G are used as the light space image. FIG. 2 is a diagram showing a light space image (evaluation target image) used in the light impression evaluation method according to the embodiment.

FIG. 2A is a photograph of the street at night in the city. The photograph of (a) includes lighting of stores, buildings, etc., car headlamps, street lights, etc. as a light source of light.

FIG. 2B is a photograph of a room in a house illuminated by a ceiling light. In the photograph of (b), the entire room is illuminated with green light instead of white light.

FIG. 2C is a photograph of the outdoors illuminated by illuminations of various colors. In the photograph of (c), the front side of the photograph is the color centered on green, the back side of the photograph is the color centered on blue, and the upper half of the photograph is the black night sky.

FIG. 2D is a photograph of a plurality of fireworks launched in the night sky. Fireworks include various colors centered on red, blue and green.

FIG. 2 (e) is a photograph of the moon rising in the night sky and the shore of the moonlit night when the moon is projected on the surface of the lake. The night sky and lake surface are blue as a whole.

(F) is a photograph of the exterior around the entrance of the house. In the photograph of (f), as a light source of light, outdoor lighting such as entrance lighting or in-house lighting (leakage light) such as a ceiling light is included. The light in the photograph of (f) is warm white light.

(G) is a photograph of a residential room illuminated by a ceiling light and an indirect lighting (desk stand). In the photograph of (g), both the ceiling light and the desk stand are irradiated with white light.

As described above, in the image presentation step S1, all of the seven photographs shown in FIG. 2 are presented to each user as the light space image, but the number of the light space images presented to the user is limited to seven. is not it. Further, the type of the light space image presented to the user is not limited to the photograph shown in FIG.

[Correlation coefficient calculation process]
After the image presentation step S1, the correlation coefficient calculation step S2 is performed. In the correlation coefficient calculation step S2, a plurality of predetermined impression evaluation words and a preference evaluation word representing one predetermined preference are used, and the phase relationship between the user's preference and impression for the light space image is used. Calculate the number.

In the present embodiment, as shown in FIG. 3, as an impression evaluation word, an evaluation word for nine predetermined impression items is used. Each impression item includes three sensitivity words as impression evaluation words (evaluation words).

Specifically, the first impression item includes three sensibility words, "soft", "nagomu", and "warm" as impression evaluation words.

The second impression item includes three sensibility words, "faint", "pale", and "tasteful", as impression evaluation words.

The third impression item includes three sensitivity words, "relieved", "eye-friendly", and "comfortable", as impression evaluation words.

The fourth impression item includes three emotional words, "clear", "wake up", and "sharp" as impression evaluation words.

The fifth impression item includes 33 sensitivity words of "lively", "lively", and "active" as impression evaluation words.

The sixth impression item includes three sensitivity words, "gorgeous", "presence", and "luxury", as impression evaluation words.

The seventh impression item includes three sensibility words, "refreshing", "transparent", and "refreshing", as impression evaluation words.

The eighth impression item includes three sensitivity words, "romantic," "elegant," and "fantastic," as impression evaluation words.

The ninth impression item includes three sensibility words, "deep", "shaded", and "spread", as impression evaluation words.

Further, as shown in FIG. 4, as a preference evaluation word expressing a person's preference, an evaluation word for one predetermined favorable item is used. This liking item includes three sensibility words, "nice", "beautiful", and "favorite", as preference evaluation words (evaluation words).

It should be noted that each sensitivity word in each of the above 10 items is obtained by collecting and classifying in advance as evaluation words related to various lights.

Then, in the correlation coefficient calculation step S2, each user's preference for the seven photographs (light space images) shown in FIG. 2 is used by using such nine items of impression evaluation words and one item of preference evaluation words. Calculate the correlation coefficient between impression and impression.

Specifically, first, the user performs a graded evaluation of a total of 10 evaluation words, 9 items of impression evaluation words and 1 item of preference evaluation words, for the 7 photographs presented in the image presentation step S1. Let me do it. That is, the user determines the degree of fit for the evaluation word of each of the 10 items for each photograph in a plurality of stages.

In this embodiment, a 7-point evaluation was performed using the Likert scale. For example, for each of the seven photographs, as shown in FIG. 5, the user has "very applicable", "applicable", "somewhat applicable", "neither" for each of the 10 items. A seven-level evaluation of "somewhat not applicable", "not applicable", and "very not applicable" is performed. The user selects an impression of each photograph from seven options for each of the ten items.

In this way, it is possible to search for the mechanism of the difference in preference between users for the light space by performing a graded evaluation between the preference and impression of each user for the light space image.

When performing the graded evaluation of each photograph, the user evaluates without distinguishing between the impression evaluation word of 9 items and the preference evaluation word of 1 item. That is, the user is made to evaluate all 10 items as items in the same row. Therefore, as shown in FIG. 5, the preference evaluation word is mixed in the impression evaluation word so as not to make the user recognize that it is the preference evaluation word.

In addition, such evaluation is performed by a plurality of users. In the present embodiment, 600 users are asked to evaluate each photograph, but the number of users is not limited to 600. The order of the 10 items shown in FIG. 5 may be changed for each user.

Then, based on the evaluation result of each user for the seven photographs, the correlation coefficient between the preference and the impression of each user for the seven photographs is calculated. That is, the strength of the relationship between the intensity of the preference (favorability) felt for the light space image and the intensity of the impression felt for the light space image in each user is defined as one preference evaluation word (favorability item) and 9. Calculated as a correlation coefficient with one impression evaluation word (impression item). Hereinafter, a specific description will be given.

The correlation coefficient represents the strength of the relationship between two variables. In this embodiment, Pearson's correlation coefficient (product-moment correlation coefficient), which is widely used as an index of the strength of the linear relationship between two variables, is used as the correlation coefficient. Pearson's correlation coefficient is an index that takes a value in the range of -1 to 1, and the closer it is to 1, the higher the positive correlation, while the closer it is to -1, the higher the negative correlation. Also, the closer it is to 0, the more uncorrelated it is.

は、それぞれx，yの平均である。 Here, assuming that the two variables for which the relationship is to be obtained are x and y, the data of n users x = (x ₁ , x ₂ , ..., X _n ), y = (for the two variables x and y. When y ₁ , y ₂ , ..., y _n ) are obtained, Pearson's correlation coefficient R is expressed by the following (Equation 1). In (Equation 1),

Is the average of x and y, respectively.

Using this (Equation 1), the correlation coefficient is calculated from the experimental results of impression evaluation of 600 users. Specifically, the data obtained by each user's 7-step impression evaluation for 7 photographs can be viewed as data of 7 stimuli × 10 items when viewed for each user. Therefore, for example, the correlation coefficient R between the favorable impression item (preference evaluation word) and the single impression item (impression evaluation word) in the user A can be obtained by the above (Equation 1). By repeating this for all nine impression items (impression evaluation words), nine correlation coefficients R can be obtained for user A. By performing this process on all users (600 people), an index showing the relationship between the nine impressions and the preference (favorability) can be obtained for each user.

That is, the (i, j) element of the matrix R having a size of n × 9 obtained as a result is r _{i, j,} and the vector of length 7 indicating the data of the favorable item in the user i is x _i, and each impression is given. Assuming that a matrix having a size of (7 × 9) indicating item data is Y _i , the resulting data is represented by the following (Equation 2). In (Equation 2), x _ik is the value of the favorable item for the kth of the user i, and y _ijk is the value of the user i evaluating the kth stimulus by the impression item j.

[Cluster classification process]
After the correlation coefficient calculation step S2, the cluster classification step S3 is performed. In the cluster classification step S3, the users are classified into clusters based on the similarity of the combination of features by performing cluster analysis (clustering) using the correlation coefficient calculated in the correlation coefficient calculation step S2.

Specifically, by performing cluster analysis using the data matrix R obtained in the correlation coefficient calculation step S2, some users are selected according to the similarity of the combination of features (that is, depending on whether the correlation is high or low). Divide into small groups (clusters) and group them.

In this embodiment, Ward's method, which is a kind of hierarchical clustering, is used as the cluster analysis. By using Ward's method, users can be accurately classified into clusters with similar correlation tendencies. Specifically, hierarchical clustering is performed by the following procedure.

Step 1: First, find the inter-cluster distance for all cluster pairs.

Step 2: Next, a new cluster is created by combining the cluster pairs with the minimum inter-cluster distance.

Step 3: Next, find the distance (update distance) between the new cluster and the other clusters.

Step 4: Repeat steps 2 and 3 until the number of clusters reaches a predetermined number (usually 1).

は、クラスタiの重心ベクトル（＝平均ベクトル）とクラスタjの重心ベクトル（＝平均ベクトル）の間の距離である。また、n_i及びn_jは、それぞれクラスタiおよびjに含まれるユーザの数である。 For example, for the cluster i and cluster j, cluster distance d _ij by Ward method is represented by the following equation (3). In (Equation 3)

Is the distance between the centroid vector (= average vector) of cluster i and the centroid vector (= average vector) of cluster j. Also, n _i and n _j are the number of users included in clusters i and j, respectively.

Since it is considered that three factors are involved in the impression evaluation of the light space, it is assumed that there is a small group related to each factor, and the number of clusters is set to 3 in this embodiment.

Therefore, as a result of performing the cluster analysis as described above, all the users who evaluated the seven photographs (light space images) were cluster 1 (first cluster), cluster 2 (second cluster), and cluster 3 (cluster 3). It is classified into three clusters (third cluster).

FIG. 6 shows the relationship between the user's correlation coefficient calculated by the light impression evaluation method according to the embodiment and the classified clusters. FIG. 6 shows the correlation coefficient between the nine impression evaluation words (impression items) and the preference evaluation words (favorable items) for each cluster classified into the three clusters of cluster 1, cluster 2, and cluster 3. Has been done. The value of the correlation coefficient shown in FIG. 6 is an average value of users belonging to each cluster. Further, in FIG. 6, the item of the impression evaluation word having a relatively large absolute value of the correlation coefficient in each class is surrounded by a broken line.

As shown in FIG. 6, for cluster 1, the preference evaluation words are "clear / awake / sharp", "lively / lively / active", and "brilliant / presence". There is a strong correlation with the impression evaluation words that indicate the activity (high energy awakening) of "a certain / luxurious" and "romantic / elegant / fantastic".

On the other hand, for cluster 2, the preference evaluation words are "soft / relaxing / warm", "faint / pale / tasteful", "relieved / eye-friendly / comfortable", and "refreshing / transparent". There is a high correlation with the impression evaluation words that indicate low tension awakening such as "yes / refreshing", "romantic / elegant / fantastic" and "deep / shaded / spread".

Further, regarding cluster 3, the preference evaluation word has a high correlation with all the impression evaluation words.

As described above, in the present embodiment, the users are classified by performing the cluster analysis based on the calculated correlation coefficient of each user. That is, users are classified based on the relationship between two factors, preference and impression. As a result, the clusters in which users are classified include cluster 1 (first cluster) in which high energy arousal is associated with preference (preference) and cluster 2 in which low tension arousal is associated with preference (preference). It was found that (second cluster) and cluster 3 (third cluster) in which both the items related in cluster 1 and the items related in cluster 2 are involved are included.

The difference between each cluster can be interpreted as Thayer's mood two-factor model. Thayer proposes the theory that a person's mood consists of tension arousal, which indicates the degree of tension, and energy arousal, which indicates the degree of activity. Therefore, the above three clusters (cluster 1, cluster 2, cluster 3) fit Thayer's mood two-factor model.

(Example)
Next, a specific example of the light impression evaluation method in the above embodiment will be described with reference to FIG. 7. FIG. 7 is a schematic diagram of an example of the light impression evaluation system 1 for executing the light impression evaluation method according to the embodiment.

As shown in FIG. 7, in the light impression evaluation system 1 in this embodiment, the light impression evaluation method was executed using the terminal device 10 and the server 20.

The terminal device 10 is, for example, an information processing device such as a general-purpose personal computer, and has a main body unit 11, a display unit 12, and an input unit 13. The display unit 12 and the input unit 13 are connected to the main body unit 11.

The main body 11 has a storage unit including an arithmetic processing unit (control unit) such as a CPU and a volatile or non-volatile semiconductor memory or the like.

The display unit 12 is a display having a display screen. The display unit 12 has a light space image (photograph) shown in FIG. 2, a menu for the user to evaluate (for example, an evaluation input item shown in FIG. 5), and an evaluation result (for example, a phase shown in FIG. 6). The number of relationships) etc. is displayed.

The input unit 13 is a user interface such as a keyboard and a mouse, and inputs various information to the main body unit 11. The impression evaluation determined by the user is input to the input unit 13 by the input operation of the user. For example, the user operates the input unit 13 while looking at the photo (light space image) displayed on the display unit 12, and inputs the evaluation for each evaluation word to the terminal device 10 for each photo.

Instead of the display unit 12 and the input unit 13, a touch panel type display or the like having the functions of both the display unit 12 and the input unit 13 may be used. For example, the terminal device 10 may be a mobile terminal such as a smartphone.

The server 20 is connected to the terminal device 10 via a communication line such as a computer network. The server 20 transmits information to the terminal device 10 and receives information from the terminal device 10.

The server 20 is, for example, an information processing device such as a computer, and includes an arithmetic processing unit that performs various arithmetic processes such as calculating the correlation coefficient R and performing cluster analysis based on the information transmitted from the terminal apparatus 10. Have.

Further, the server 20 has a storage unit that stores in advance the light space image (photograph) shown in FIG. 2, the evaluation words shown in FIGS. 3 and 4, information related to the stage evaluation shown in FIG. 5, and the like. Information obtained by executing the light impression evaluation method (information input by the user, processing result of the arithmetic processing unit, etc.) may be stored in the storage unit. The storage unit may further store other information, or may store a series of programs and the like required for processing for executing the light impression evaluation method.

Next, a specific procedure when the impression evaluation method of the light is performed by the impression evaluation system 1 will be described.

In this embodiment, the seven photographs shown in FIG. 2 were evaluated by 600 users. Specifically, 150 people were evaluated in each age group of "20's and 30's", "40's", "50's", and "60's and over".

Specifically, first, the light space image is transmitted from the server 20 to the terminal device 10 and displayed on the display unit 12, so that the light space image is presented to each user (image presentation step). In this embodiment, seven photographs shown in FIG. 2 as light space images are sequentially displayed on the display unit 12.

Next, the correlation coefficient between each user's preference and impression for each light space image is calculated using 10 items of evaluation words of 9 items of impression evaluation words and 1 item of preference evaluation words (phase). Relationship number calculation process). In this embodiment, as shown in FIG. 7, the user evaluates each impression of 10 items (see FIG. 5) received from each photograph while looking at the photograph displayed on the display unit 12, and evaluates the impression. The result is input by the input unit 13. Specifically, on the menu screen (evaluation selection screen) shown in FIG. 5, the user selects an impression applicable to each of the 10 items from the seven options, and selects the corresponding option with the mouse (input unit 13). Move the cursor and click. As a result, the impression evaluation of the user is input. The result of the user's impression evaluation input to the input unit 13 is transmitted to the server 20. Based on the result of the user's impression evaluation, the server 20 calculates the correlation coefficient of the nine impression evaluation words with respect to the preference evaluation word using the above (Equation 1) and (Equation 2).

Next, all users are classified into clusters based on the similarity of the combination of features by performing cluster analysis using the calculated correlation coefficient (cluster classification step). In this embodiment, cluster analysis is performed on the server 20. Specifically, cluster analysis was performed by the Ward's method using the above (Equation 3). As a result, all users are classified into three clusters, cluster 1, cluster 2, and cluster 3, and as shown in FIG. 6, the correlation coefficient of nine impression evaluation words with respect to the preference evaluation word is obtained for each cluster. rice field.

(summary)
As described above, the light impression evaluation method in the present embodiment includes a step of presenting an image including a light state to the user, a plurality of predetermined impression evaluation words, and a preference evaluation word representing one predetermined preference. The process of calculating the correlation coefficient between each user's preference and impression for the image including the light state using, and the cluster analysis using the calculated correlation coefficient are used to combine the characteristics of the user. Includes the process of classifying into clusters by similarity.

As described above, in the light impression evaluation method in the present embodiment, a plurality of impression evaluation words and one preference evaluation word are used, and the preference and impression of each user for the image including the light state are used. After calculating the correlation coefficient, cluster analysis is performed using this correlation coefficient. This makes it possible to clarify the relationship between the impression and preference of each user, and then classify the users based on the relationship. That is, it is possible to perform grouping based on the difference in how the users feel the favorable impression.

Therefore, even if there is diversity among users in the relationship between the impression given by the light and the preference of the user, the impression of the light received by a person can be accurately evaluated. This makes it possible to build a light space that many people like. As a result, based on this evaluation, it is possible to easily design and plan a design related to lighting and the like, or design a lighting space and the like. For example, the design time for bringing the lighting space closer to the ideal state can be shortened.

Also, in the present embodiment, the clusters are a first cluster in which high energy arousal is associated with preference, a second cluster in which low tension arousal is associated with preference, a first cluster and a second. Includes a third cluster, which involves both of the clusters of.

As described above, the light impression evaluation method in the present embodiment is suitable for the mood two-factor model.

Further, in the present embodiment, the cluster analysis is Ward's method.

This makes it possible to accurately classify a plurality of users into clusters having similar correlation tendencies.

(Modification example)
Although the method for evaluating the impression of light according to the present invention has been described above based on the embodiment, the present invention is not limited to the above embodiment.

For example, in the above embodiment, nine items of impression evaluation words are used, but the impression evaluation words are not limited to nine items.

Further, in the above embodiment, the number of Kansei words included in the impression evaluation word and the preference evaluation word is three, but the number of each item of the impression evaluation word and the Kansei word included in the preference evaluation word is three. The number is not limited to one, and may be four or more. That is, the impression evaluation word and the preference evaluation word need only include at least three sensitivity words. Further, the specific words of the impression evaluation word and the sensitivity word of the preference evaluation word are not limited to those disclosed in the above-described embodiments and examples.

In addition, it can be realized by subjecting various modifications to the above-described embodiment to those skilled in the art, or by arbitrarily combining the components and functions of the above-described embodiment without departing from the spirit of the present invention. The form to be used is also included in the present invention.

Further, in the above description, the control unit such as the arithmetic processing unit in the terminal device 10 and the server 20 may be a circuit. These circuits may form one circuit as a whole, or may be separate circuits from each other. Further, each of these circuits may be a general-purpose circuit or a dedicated circuit.

Further, the light impression evaluation method in the above embodiment may be executed by one computer as an impression evaluation device. In this case, the impression evaluation device has the configurations and functions of the terminal device 10 and the server 20 described above, and the computer executes a program using hardware resources such as a processor (CPU), a memory, and an input / output circuit. By doing so, the processing of each step of the light impression evaluation method is executed. Specifically, the processor or the like acquires the data to be processed from the memory or the input / output circuit or the like and calculates the data, or outputs the calculation result to the memory or the input / output circuit or the like to process each process. To execute.

Further, the program for executing the processing of each step of the light impression evaluation method in the above embodiment may be recorded on a non-temporary recording medium such as a computer-readable CD-ROM. In this case, the computer reads the program from the non-temporary recording medium and executes the program to execute the processing of each step.

Further, the present invention can be realized as a program for executing the light impression evaluation method in the above-described embodiment, or can be realized as a computer-readable recording medium in which the program is stored.

1 Light impression evaluation system 10 Terminal device 11 Main unit 12 Display unit 13 Input unit 20 Server

Claims (3)

A method of evaluating the impression of light performed by a computer.
The process of presenting an image including the lighting state to the user,
Based on the evaluation results obtained by performing a stepwise evaluation of each user's preference and impression of the image using a plurality of predetermined impression evaluation words and one predetermined preference evaluation word. A step of calculating the correlation coefficient between the one preference evaluation word and the plurality of impression evaluation words for each user, and
It includes a step of classifying users into clusters based on the similarity of feature combinations by performing cluster analysis using the correlation coefficient.
The preference evaluation word is a sensitivity word indicating a person's liking, and is
The impression evaluation word is a sensitivity word indicating an impression that a person receives.
Light impression evaluation method. The clusters are a first cluster in which high energy arousal and the preference are associated with a high degree of mood activity, and low tension arousal and the preference, which indicate a low mood tone. Includes a second cluster associated with, and a third cluster involving both the first cluster and the second cluster.
The method for evaluating the impression of light according to claim 1. The cluster analysis is Ward's method.
The light impression evaluation method according to claim 1 or 2.

Images