JPH09204138A - Method and device for forming hair color chart - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately display finished dyeing color according to the nature of hair by a method and device for hair color chart formation which enable the finished dyeing color to accurately be grasped before hair is colored. SOLUTION: A hair nature deciding process is performed which decides the nature of hair of a customer before the hair is colored (S1). Then a dyeing process which colors the hair of the customer (S2) is performed. A hair color deciding process which decides the color of the dyed hair is performed. A neural network analyzes the hair nature and hair color and learns the results, and based on the the learnt results, a hair color associating process which associates a finished dyeing color corresponding to the inputted hair nature is carried out (S4). A chard forming process which from a hair color chart by relating the found hair color and hair nature with each other is performed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a hair color chart making method and a hair color chart making apparatus, and more particularly to a hair color chart making method and a hair color chart making it possible to accurately grasp a color after dyeing before a hair color treatment. The present invention relates to a color chart creation device.

[0002]

2. Description of the Related Art In recent years, so-called "fashionable dyeing", in which hair is dyed brightly with a hair color to change the color of the hair, has become popular. This "fashion dyeing" is already recognized as a part of hairstyles, especially among women. Along with this,
The treatment rate for hair coloring has been increasing rapidly for about five years, and in the fall of 1994, the treatment rate has increased as one in four people have been doing hair coloring.

However, the dyed color may vary greatly depending on the hair quality, and the desired color may not be obtained.
For example, for thick and undamaged hair and thin and damaged hair, even if the same hair color is used, the change in brightness after dyeing will be different, and a difference will appear as if dyed with different hair colors.

Therefore, in a beauty salon, a professional hairdresser sensuously grasps the hair quality of a person who wants to dye hair (hereinafter referred to as a customer), and associates it with empirical knowledge to bring it closer to a desired color. Is being done. However, since communication that conveys colors in words is abstract, it is difficult to convey the color desired by the customer to the hairdresser, and there is the fact that the customer feels uneasy whether or not the color will be the same as the image.

Also, when the customer directly purchases the hair color, although the dyed-up color of the product is shown on the package, no information about the difference in the dyed-up color depending on the hair quality is shown, so the customer It is difficult to choose a hair color that suits your hair quality. To improve this,
Some retailers now have hair bundle swatches to convey the dyed color. This hair bundle sample is made up of nylon hair that has been dyed for each color number, but since it is a partial expression, it is difficult to imagine the entire hair,
Further, in the case of the hair bundle sample, the information on the hair quality condition is not considered.

Therefore, it is possible to select a color according to the hair quality of each individual customer, to easily imagine the state of dyeing, and at the time of counseling, the hairdresser and the customer can obtain a common recognition. Development of a sample is desired. Such a hair dye sample can be created by performing a monitor test and collecting and analyzing data of the hair quality of the customer and the color of the dyed hair before and after the hair dyeing operation.

[0007]

However, there are various products for hair color due to the diversification of customer needs, and a large number of colors (color numbers) are set for each product. There are various types. Therefore, if a monitor test is performed on all products and color numbers for each hair type, the number of people, the amount of money and the time spent for this test will be enormous.

When the hair quality is judged in the monitor test, this judgment is made by a professional engineer. Professional engineers intuitively grasp the color and hair quality with the five senses, which is based on many years of experience of professional engineers. Therefore, it is difficult for a professional engineer to objectively evaluate the customer's hair quality as test data.

The present invention has been made in view of the above points, and an object of the present invention is to provide a method for producing a hair color chart and a hair color chart producing apparatus in which the color of the dyed color is accurately displayed according to the hair quality. And

[0010]

[Means for Solving the Problems] The above problems can be solved by taking the following measures. Claim 1
In the invention method described above, a hair quality determination process for determining the hair quality of the customer before the hair color treatment, and a hair dyeing process in which the hair color is applied to the customer's hair after the hair quality determination process is finished. Treatment, a hair color determination process for determining the hair color of the hair dyed in the hair dyeing process, a hair quality determined by the hair quality determination process, and a hair color determined by the hair color determination process. A hair color associating process for associating the dyed color corresponding to the hair quality input based on the learning result with the neural network by analyzing with a neural network and learning the result, and the input hair quality ,
And a chart creation process for creating a hair color chart by associating with a dyed color corresponding to the hair quality associated by the hair color association process.

According to the invention method of claim 2, in the method of creating a hair color chart according to claim 1, the hair quality judgment process is performed by a plurality of judges, and the neural network is used by the judges. It is characterized in that the results are analyzed by preparing for the number of persons and in the hair color association processing, based on the hair quality judged by each judge and the hair color of the dyed hair.

Further, in the invention according to claim 3, the hair quality data obtained by judging the hair quality of the customer before the hair color treatment and the hair for the customer's hair after the hair quality judgment processing are completed. Input means for inputting hair color data obtained by determining the hair color of the dyed hair after color treatment and hair dyeing, and hair quality data and hair color data input from the input means Is analyzed by a neural network to learn the result, and a hair color association means for associating a dyed color corresponding to the hair quality input based on the learning result, and a hair quality input from the input means. A hair color chart output means for outputting a hair color chart in which data is associated with a dyed color corresponding to the hair quality associated with the hair color association processing. It is an feature.

The above means operate as follows. According to the first and third aspects of the present invention, first, the hair quality of the customer is judged before the hair color treatment, and then the hair color is applied to the customer's hair and the hair is dyed. Determine the hair color of. Further, the determined hair quality and hair color are analyzed by a neural network and the result is learned. Therefore, this learning result becomes learning data in which the hair quality and the hair color after the treatment are correlated.

Therefore, it is possible to estimate the dyed hair color by inputting the hair quality based on this learning data. That is, the neural network is a concept of information processing using a human brain as a model, and its features are learning ability and associative ability.

The learning ability means that a causal relationship of data can be stored by making the best use of the characteristics of the data. Here, an intuitive evaluation is objectively made by incorporating the experience and knowledge of the judge to determine the relationship between the hair quality and the dyed color. It is possible to analyze by including the knowledge obtained empirically. On the other hand, associative ability is to derive an optimal answer by analogy with a small amount of data.

Accordingly, even if all the data regarding the hair quality (for example, various items such as thickness and hardness) and the hair color after the treatment are not obtained, the learning can be performed by inputting the hair quality. It is possible to associate the dyed hair color based on the learned data. Therefore, it becomes possible to evaluate the post-treatment stain with a small amount of input data (small-scale monitor test), reduce the number of monitors, and improve the efficiency and cost of the monitor test.

Further, since the hair color chart is displayed in association with the hair quality and the dyed color obtained as described above, the customer can easily understand the dyed state. According to the invention of claim 2, the hair quality determination process is performed by a plurality of determiners, the neural networks are prepared for the number of the determiners, and the hair quality determined by each determiner in the hair color association process is determined. By configuring the learning result to be analyzed based on the hair color of the dyed hair, it is possible to disperse individual differences among each judge and generate highly accurate learning data. Is possible.

[0018]

BEST MODE FOR CARRYING OUT THE INVENTION Next, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a process chart showing a method for producing a hair color chart, which is an embodiment of the present invention. As shown in the figure, the method of creating a hair color chart according to the present embodiment is a hair quality determination process of determining the hair quality of a customer and creating hair quality data before the hair color treatment (step 1, in the figure). S1), and a hair dyeing process (step 2) in which hair color is applied to the customer's hair after the hair quality determination process is finished, and the hair color of the hair dyed in the hair dyeing process is changed. A hair color determination process (step 3) for determining and creating hair color data, and an analysis process for analyzing the hair quality determined by the hair quality determination process and the hair color determined by the hair color determination process (step 4) ) And a chart creating process (step 5) for creating and outputting a hair color chart based on this analysis process.

In the present embodiment, in the analysis process, the hair quality determined in the hair quality determination process and the hair color determined in the hair color determination process are analyzed by a neural network to learn the result, and this learning is also performed. It is characterized in that a dyed color corresponding to the hair quality input based on the result is associated with the neural network (hereinafter, the processing of step 4 is referred to as hair color association processing).

FIG. 8 shows an embodiment of a hair color chart creating apparatus 1 for carrying out the above hair color chart creating method. The hair color chart creating apparatus 1 is composed of a personal computer 2, a display device 3, a printer 4, an interface device 5, and the like.

The personal computer 2 is provided with a keyboard 6 (which functions as an input means) for inputting the hair quality data created in step 1 and the hair color data created in step 3 described above. Further, the hair color associating process executed in the above step 4 is configured as software which is stored in the storage device of the personal computer 2 and executed by the arithmetic processing unit.

In this sense, the personal computer 2 analyzes the hair quality data and the hair color data input from the keyboard 6 (input means) by a neural network and learns the result, and inputs based on this learning result. It functions as a hair color associating means that associates a dyed color corresponding to the hair quality.

Further, the display device 3 and the printer 4
Outputs a hair color chart in which the hair quality data input from the input means is associated with the dyed color corresponding to the hair quality associated with the hair quality data. When it is output to the display device 3, it is output as image information, and when it is output to the printer 4, it is output as a paper surface.

Next, a method for creating a hair color chart actually performed by the inventor using the hair color chart creating apparatus 1 configured as described above will be described below. As described above, in order to create the hair color chart, first, the hair quality determination process of determining the hair quality of the customer and creating the hair quality data is performed before the hair color treatment. The data collection of hair quality was performed as follows.

Judges: 5 professional technicians who perform hair color treatments Panels (customers): 15 people each with a combination of presence or absence of hair dye / perm 60 people Further, in this embodiment, the dyed color In order to capture the change as an effect of only the hair quality, the judgment items for the hair quality were set as follows. [Hair quality items] ・ Thickness (thick, slightly thick, ordinary, slightly thin, thin) ・ Hardness (hard, slightly hard, ordinary, slightly soft, soft) ・ Damage (damaged, slightly damaged, no damage) -Hair dye (yes / no) -Perm (yes / no) The above items were judged by sensory evaluation based on the visual and tactile sensations of the judge. As a result of the determination, a hair color check paper 7 as shown in FIG. 2 is used, and each determination result is, for example, thin 1, slightly thin 2, ordinary 3, ... Taking the determination of the thickness of the hair as an example. It is expressed as a numerical value and input to the personal computer 2.

When the hair quality determination process for determining the hair quality performed before the hair color treatment is completed as described above, a hair dyeing process for performing hair color treatment and hair dyeing is subsequently performed.
The conditions for hair dyeing treatment are as follows.・ Used color: Oxidative hair dye (Product name: Zotosu Rouge 7Y) ・ Treatment method: Same as normal hair color ・ Left time: 20 minutes after application of color judgment area ・ Environmental conditions: environment such as temperature and humidity In addition, the used color Zotos Herouge is the product with the most colors, the 7Y color number is popular, and it is located in the middle of the fashionable color classification. Selected from.

When the hair dyeing process is completed as described above, subsequently a hair color judging process for judging the hair color of the dyed hair is carried out. The hair color was judged by judging the color tone and the brightness. Specifically, the hair color determination items are as follows. [Color tone] The color tone of the hair is shown, and is referred to as color here, and in the present embodiment, the following eight colors are used.

-Wine (W) Reddish with a slight bluish-Red (R) Reddish-Orange (Or) Reddish with a slight yellowishness-Normal (Nm) Natural color that does not belong to R or Or-Yellow (Y ) Yellowness ・ Gleen (Gr) Greenness ・ BlueGreen (Bg) Slightly bluish green ・ Blue (Bl) Blueness [Brightness] Indicates the degree of lightness of the hair color. Brightest tones from 9
Use up to.

For the color determination, a semi-long, slightly left-sided photograph of the upper part of the neck of a woman was used as computer graphics, and a tool with a visual expression in which the brightness of the color was changed was used. Based on this, the color of the hair is judged by the tool that created the photograph processed for each color and tone.

In the present embodiment, one tool (color map) in which the above eight hues are changed to hair color and one in which the brightness is changed within the same hue (tone map) are used.
Was prepared for each color (8 sheets), and the hair color was judged using this color map and tone map.

When the hair color determination process is completed as described above, the hair color after dyeing corresponding to the hair quality is identified. This hair quality data and hair color data corresponding to the hair quality are
The operator checks the hair color check sheet 7 shown in FIG.
The user operates the keyboard 6 while looking at the input to the personal computer 2.

When the input processing of the hair quality data and the hair color data corresponding to the hair quality is completed as described above, the personal computer 2 analyzes the input data. By the way, considering the combination of the evaluation items for the hair quality items used this time, there are 300 types, and if we try to obtain all the data in the actual monitor test, the number of panels (customers) required will be at least 300. Also, considering the types of hair colors, the number of combinations further increases, and a huge number of people, cost and time are required. Further, even if all the data are obtained, the habits of each judge are included, and therefore, it is necessary to take this into consideration when performing the evaluation.

Therefore, in the present embodiment, as an analysis method that enables analysis of all combinations with a small amount of data (data of 60 persons as described above), an analysis method using a neural network is adopted. Characterize. As a result, the efficiency of the monitor test can be significantly improved, and human experience and knowledge can be incorporated as know-how.

Here, the neural network will be described. The neural network is a concept of information processing using a human brain as a model, and in this embodiment, a hierarchical neural network as shown in FIG. 3 is adopted. Further, the features of the neural network are learning ability and associative ability as shown in FIG.

The learning ability means that a causal relationship of data can be stored by making the best use of the characteristics of the data. The characteristics of the data in the present embodiment include the experience and knowledge of the judge because the relation between the hair quality and the dyed color is based on the judgment result judged by the judge without using a mechanical method. It is possible to perform such analysis.

As a concrete learning process, for example, "thickness → thinness", "damage → present", and "perm →
In the case of “none”, the hair color is “color → red”,
If "tone → 9", the hair quality data and the hair color data are associated with each other and stored as a learning pattern.

On the other hand, the associative ability is to derive an optimum answer by analogy with a small amount of data. Even if all combinations of hair quality items are not obtained by this, all combinations (300 data) from the existing data (data for 60 people).
Data can be estimated (association), and all combinations can be evaluated simply by conducting a small-scale monitor test, which can reduce the number of monitors and improve the efficiency of analysis processing.

As a concrete associative process, when “thickness → normal”, “damage → present”, and “perm → none” are input as hair qualities, even if there is no matching hair color data. When a neural network is used, a difference can be seen in the item of "thickness", but its influence is small, and "color → red" and "tone → 9" are associated and output as hair color data.

FIG. 5 shows a flow chart of the hair color associating process executed by the personal computer 2 based on the above items. When the analysis process shown in the figure is activated, in step 10 (the step is abbreviated as S in the figure), a determination process is performed for each item of hair quality data and hair color data input from the keyboard 6. This determination process is performed for each item of hair quality data and hair color data. In the hair color data, the colors are digitized such as wine: 1, red: 2 ... blue8.

In step 10, a process for associating the hair quality data with the hair color data is also performed. In the illustrated example, "thickness → 1 (thin)" and "hardness → 4 (somewhat soft)". The hair quality data of "," damage → 2 (some damage), "hair dye → 1 (yes)", "perm → 1 (yes)" are "color → 2 (red)", "tone → 8"
Is associated with the hair color data as a learning pattern. In this embodiment, the following items are set as rules for generating learning patterns.

(1) Collect the hair quality data for each dye and tone after the treatment (2) Combine the same hair quality pattern in the same dye and tone classification (3) Other dye The process of (1) is also performed for the classification of colors and tones. (4) The same hair quality pattern is used for different classifications of different dyed colors and tones. The rules for generating learning patterns are as described above. By setting, the learning patterns generated by the contents of the hair quality data are collected, and the number of learning patterns may be smaller than the number of monitors (60 in this embodiment).

The learning pattern thus subjected to the association processing in step 10 is input to the neural network in step 12 and learning is performed.
In this embodiment, the back propagation method is used for this learning. In the following step 14, the processes of steps 10 and 12 described above are repeatedly performed for each judge (five persons in this embodiment). Further, in step 16, a network is configured for each judge, and in step 18, each network constructed in step 16 is connected to each network.
Enter all 00 types of hair quality patterns (In addition, of the hair quality patterns,
In the present specification, it is particularly referred to as hair quality data).

In the following step 20, the learning patterns output for each network are integrated. At this time, a hair quality pattern that is not stored as a learning pattern in step 18 is also input. In this case, however, a hair color that is most suitable for the input hair quality pattern is obtained due to the associative ability of the neural network described above. It is associated.
Therefore, the learning pattern output from each network in step 20 is output for all (300 types) hair quality patterns.

In the following step 22, the classification result is output based on the result integrated in step 20. As is clear from the above description, when a hair quality pattern is input to the neural network according to the present embodiment, a dyed color (hair color) is output. Therefore, by inputting all possible hair quality patterns (300 types), the output becomes a classification result in which the hair quality and the hair color are associated with each other.

As described above, in this embodiment, the hair quality determination process is performed by a plurality of determiners, neural networks are prepared for the number of the determiners, and the hair quality determined by each determiner in the hair color association process is determined. The result of analysis based on the hair color of the dyed hair is stored as a learning pattern. As a result, it is possible to reflect the experience and knowledge of each determiner in the learning pattern and then to vary the individual differences among the determiners, and it is possible to generate highly accurate learning data.

When the hair color association processing is completed as described above,
Subsequently, a chart creating process for creating a hair color chart is performed. In this embodiment, when the hair color chart is created, the thickness and the hardness are correlated, and therefore the thickness also includes the hardness, and the hair color chart is classified from four items of thickness, damage, hair dye, and the presence or absence of perm. It was created. It should be noted that it is difficult to think that there is no damage in the combination of either "hairy" or "perm", so it was excluded.

FIG. 6 shows a first application example of the hair color chart, and is an example in which a hair color chart is created for each color number. The hair color chart shown in the figure is
It expresses the dyeing of the hair quality for each color number, and it is possible to capture the change in dyeing due to the hair quality. As a result, the hair color chart can be used as a material for basic data at a salon, for education of new employees, and as an educational material at a store.

FIG. 7 shows a second application example of the hair color chart, which is an example in which a hair color chart is created for each hair type. The hair color chart shown in the figure expresses the dyeing by color number for each hair type, and customers can select products from the image of dyeing. This makes it possible to utilize this hair color chart as a counseling tool for customers at the store and as a catalog for self-sale.

The hair color charts shown in FIGS. 6 and 7 are output from the color printer 4 (see FIG. 8), but the output form of the hair color chart is limited to the output from the printer 4. Alternatively, the display device 3 may be used to output an image.

Next, the result of comparison between the hair color chart prepared as described above and the hair color as a result of actually applying the hair color will be described below. In this comparison process, when comparing the actual dyed hair color with the hair color printed on the created hair color chart, the following cases were determined to match both hair colors. That is, in the above-mentioned 8 color classifications and 9 tone classifications, ・ They are completely the same ・ The colors are the same and the tones are different (one dark, the one is bright) ・ The tones are the same and the colors are different (one It was assumed that the hair colors of both were the same (red, one yellow). In addition, the determination was performed by the above-mentioned 5 professional engineers.

As a result, it was found that 70% of the hair color printed on the hair color chart was in agreement with the actual dyed hair color. This is 9 for non-expert customers
This is equivalent to obtaining 0% agreement, and therefore the effectiveness of the hair color chart is considered to be high, and it can be said that the degree of agreement is sufficient.

[0052]

As described above, according to the present invention, it is possible to give the customer a sense of security in selecting the color of the hair color and to make the customer more satisfied than the hair bundle sample. It is possible to have a common recognition of the dyed hair color between the customer and the beautician and the customer.

[Brief description of drawings]

FIG. 1 is a principle diagram of a method of creating a hair color chart that is an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a hair color check paper used in a method for creating a hair color chart.

FIG. 3 is a diagram showing an example of a neural network model.

FIG. 4 is a diagram for explaining a learning ability and an associative ability of a neural network.

FIG. 5 is a diagram showing a flowchart of a hair color association process executed by a personal computer.

FIG. 6 is a diagram showing a first application example of a hair color chart.

FIG. 7 is a diagram showing a second application example of a hair color chart.

FIG. 8 is a configuration diagram of a hair color chart creation apparatus that is an embodiment of the present invention.

[Explanation of symbols]

 1 Hair Color Chart Creating Device 2 Personal Computer 3 Display Device 4 Printer 5 Interface Device 6 Keyboard 7 Hair Color Check Paper

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor, Adult Adult, 3-9-1, Nishigotanda, Shinagawa-ku, Tokyo Shiseido Beauty Science Research Institute

Claims (3)

[Claims] 1. A hair quality determination process for determining a customer's hair quality before a hair color treatment, and a hair dyeing process for performing a hair color on the customer's hair after the hair quality determination process is finished. A hair color determination process for determining the hair color of the hair dyed in the hair dyeing process, and the hair quality determined by the hair quality determination process and the hair color determined by the hair color determination process are neuralized. A hair color association process of associating the dyed color corresponding to the hair quality input based on the learning result with the neural network while learning the result by analyzing the network, the input hair quality, A hair color chart, wherein the hair color chart is associated with a dyed color corresponding to the hair quality associated with the hair color association processing. Forming method. 2. The method of creating a hair color chart according to claim 1, wherein the hair quality determination process is performed by a plurality of determiners, and the neural networks are prepared for the number of the determiners, and the hair color association process is performed. A method for creating a hair color chart, characterized in that the result of analysis is learned based on the hair quality judged by each judge and the hair color of the dyed hair. 3. The hair quality data obtained by determining the hair quality of the customer before the hair color treatment, and the hair color is applied to the customer's hair after the hair quality determination processing is finished to dye the hair. Input means for inputting hair color data obtained by determining the hair color of the dyed hair, and hair quality data and hair color data input from the input means are analyzed by a neural network. A hair color association unit that learns the result and associates a dyed color corresponding to the hair quality input based on the learning result, the hair quality data input from the input unit, and the hair color association process. And a hair color chart output means for outputting a hair color chart associated with a dyed color corresponding to the hair quality associated with the hair color chart output means. Creation device.