JP3240086B2 - Color simulator for design evaluation - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color simulator for design evaluation using computer graphics, and more particularly to a color simulator for design evaluation which displays candidate colors to be used for evaluation in accordance with inputted kansei terms.

[0002]

BACKGROUND OF THE INVENTION In design, color plays an important role along with shape. Conventionally, when determining the color of automobiles, home appliances, etc., it has been practiced to apply paint colors to the models and use them as judgment materials. For large-scale objects that are difficult to model as landscape materials (guardrails, bridges, etc.) Is based on a perspective drawing drawn by a designer as a source of judgment.

However, in the case of applying a paint color to a model, it takes a long time to produce the model, and in recent years, it is easy to change the color because of the fact that it does not look real and does not appear real in a perspective view. Display by computer graphics that can be combined with an actual photograph or the like is often used as judgment material.

[0004] It is desirable to determine the colors of landscape materials and the like in accordance with the surrounding scenery and locality (eg, educational district, downtown, etc.) even if the shapes are the same.

[0005] In addition, in the case of home appliances and the like that are arranged in a room, the colors appear different depending on the type of illumination (incandescent light bulb, fluorescent light, etc.), so it is necessary to consider the type of illumination. This is because the colors of automobiles and the like may look different depending on the climate of the region where they are exported, and it is necessary to check the daytime and nighttime conditions of the landscape materials. In any case, display considering the type of light source is required.

In any of the above cases, the candidate colors are determined by the subjectivity of the designer,
In a display system using computer graphics, a display simply according to an input numerical value is performed.

[0007]

In the above-mentioned system using the conventional computer graphics, the display is performed by simply inputting numerical values, and the candidate colors are determined by the designer's subjectivity. There is a problem in that the display is not necessarily appropriate, and display in consideration of the type of illumination cannot be performed.

The present invention has been made in view of the above-mentioned problems of the prior art, and is a design evaluation color simulator capable of determining and displaying candidate colors in consideration of a use environment. The purpose is to realize.

[0009]

The design evaluation color simulator of the present invention is a design evaluation color simulator for displaying candidate colors to be used for evaluation according to sensibility terms. Based on the results of a survey of “colors that feel fun” and “colors that feel light” by a large number of unspecified large numbers of humans, a sentiment database group consisting of a plurality of databases in which sentiment terms and colors are associated based on the results, An input unit for performing an input or the like for selecting a database from the sensitivity database group, an input layer corresponding to a sensitivity term, and a neural network including an output layer corresponding to a plurality of candidate colors; Display color determining means for determining a candidate color to be used for evaluation in accordance with the input content of Performs weighting network constituting the neural network in accordance with the contents stored in the selected database Te, and determining the candidate color used for evaluation using the neural network.

In this case, a conversion means for converting the candidate color determined by the display color determination means in accordance with an instruction input indicating the type of the light source input to the input means may be provided.

[0011]

In the present invention, the weight of the neural network for determining the candidate color is determined according to the storage contents of the selected database, so that the displayed candidate color is changed depending on the selected state of the database. In the present invention, each database is a database quantifying the sensibility of an unspecified number of humans such as "color that feels fun" and "color that feels light", and is further classified by region, gender, age, and the like. Therefore, it is possible to give the determined candidate color a regionality and reflect the user's preference.

When a conversion means for converting a candidate color in accordance with an instruction input indicating a type of a light source is provided, when a light source type indicating a characteristic of an area or a room is input, the candidate color to be displayed changes accordingly. Therefore, it is possible to confirm subtle colors.

[0013]

Next, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a diagram showing a functional configuration of a main part of an embodiment of a color simulator for design evaluation according to the present invention, and FIG. 2 is a diagram showing a functional configuration showing a process of generating and coloring a model in the present embodiment. It is.

As shown in FIG. 2, the generation of the display model in the present embodiment is performed by creating a wire frame based on a design image and performing a modeling process S201 for creating a skeleton. Rendering process S20 to create video with color and shading
2. With respect to the video produced in the rendering process S202, the process is performed by a painting process S204 for performing color change, mixing, and image synthesis. Since each of these processes is performed by a conventionally known technique, the description is omitted.

The library making process S203 includes a color selection from the sensibility terms characteristic of the present invention, a simulation of the color of the display model which changes depending on the place where the color is arranged, a metal texture parameter library for representing the surface material of the display model, It consists of a pattern deformation module for deforming the pattern to be displayed in accordance with the perspective, and a frame capture tool for animation for frame-forming the formed image into a moving image. Images are created by utilizing these libraries.

The configuration shown in FIG. 1 shows a portion of the library making process S203 for selecting a color from the sensibility term. Hereinafter, the configuration and operation will be described.

In this embodiment, a display color determining means 101, a paint color database 102, a sensitivity database group 103 composed of a plurality of databases, an input means 104, a Munsell color-R, G, B conversion means 105 and a display system 106.
It is composed of

The input means 104 receives an external key (not shown) input, and includes, as its contents, a numerical value indicating the shape of the display model, a database used in the sensitivity database group 103, and the importance of each database. Words indicating sensibilities such as "fun" and "light" and instruction inputs such as a light source type at a place where the display model is arranged are output to the display color determining means 101.

Each of the databases constituting the sensibility database group 103 is a database obtained by investigating sensitivities of unspecified large numbers of humans regarding colors such as "colors that feel fun" and "colors that feel lightly", and quantifies the results. And
Each database is divided into regions, genders, ages, and the like. The paint color database 102 stores colors displayed in a modified Munsell color system (hereinafter simply referred to as Munsell color) defined in JIS in association with “words indicating sensitivity”.

The display color conversion means 101 includes an input means 104
The display color is determined on the basis of the database in the sensitivity database group 103 indicated by the input contents and the storage contents of the paint color database 102.

Munsell color-R, G, B conversion means 105
Converts the Munsell color determined by the display color determining means 101 into RGB signals for display on a CRT, and the display system 106 performs display using the converted RGB signals.

The above-mentioned components are controlled by a control device (not shown), which controls the operation of the entire design evaluation color simulator of the present embodiment. , Rendering processing S202, library making processing S203
And a painting process S204.

FIG. 3 is a flowchart showing a control operation of the control device when a display image is generated from an instruction input from the input means 104, and FIG. 4 is provided in the display color determining means 101 to determine a display color. FIG. 2 is a diagram illustrating a configuration of a neural network. Hereinafter, a process of generating a display image will be described with reference to FIGS.

When the user of the apparatus inputs an instruction to determine the display color together with the database to be used, its importance, words indicating sensibility, instruction of the type of light source, and the like, the control apparatus executes: Weighting is applied to the neural network network that determines the display color according to the selected database and its importance (step S1).

The neural network of the present embodiment is of a hierarchical type, and as shown in FIG.
Provided between the input layer and the output layer to adjust the ambiguity of the output, and the input layer corresponding to the words indicating sensibility such as "thick" and "cool", the output layer corresponding to the Munsell color It is composed of an intermediate layer. The input layer is 33 units (equivalent to 33 words), the intermediate layer is 40 units, and the output layer is 28
0 units (for 280 paint colors) are provided, and each input layer and each intermediate layer and each intermediate layer and each output layer are individually connected to each other by a network, and the control device adjusts the weight of the connection of each unit. I do. In this embodiment, back propagation (back propagation method) is used as a weighting method. Note that the number of units in the output layer, the intermediate layer, and the output layer naturally changes according to the number of sensible terms and colors.

The control device inputs a signal to a unit corresponding to the inputted kansei term among the units of the input layer of the neural network weighted as described above,
The paint color is determined (step S2).

The input signal is converted by a transmission condition that changes depending on the weight of the connection, the contents of each unit in the output layer are determined, and the control device determines the display color from the value of each output unit. In determining the display color, only the one with the largest (or smallest) numerical value is displayed as the optimal candidate, and all the combinations in which the combination of a plurality of colors above (or below) a predetermined threshold value changes are ranked. May be displayed.

Next, the controller converts the paint color determined in step S2 into RGB signals by the Munsell color-R, G, B conversion means 105 in order to display the color in computer graphics (step S3). Further, the display color determination unit 101 is caused to perform correction according to the type of the input light source, and is displayed on the display system 106 (step S4).

In the conversion from Munsell colors to RGB signals in step S3, colors that do not match the measured data were converted using spline interpolation.

The Munsell color system is represented by a combination of two cones having lightness in the height direction, saturation in the radial direction, and hue in the circumferential direction. And no physical basis. Therefore, in the present embodiment, the conversion is performed based on the actually measured values specified in JIS Z8721.

The example shown in FIG. 5A shows an example in which the points shown in FIG. First, a point defined by JIS Z8721 near the point to be interpolated is obtained, and each obtained point is interpolated in the saturation direction. What is required at this time is: Next, interpolation is performed in the hue direction using the points. What is required at this time is: Subsequently, the points are interpolated in the brightness direction to obtain points.

The above interpolations are all performed using splines. The number of data points of neighboring points required for performing interpolation depends on the order of the spline, and when performing an n-order spline transformation, n + 2 data points are required.

Using the spline function curve obtained in this way, x, y chromaticity coordinate values are obtained from the x, y modified chromaticity coordinates corrected to Munsell chroma as shown in FIG. 5B. .

Next, the tristimulus values XYZ are obtained by substituting the x, y chromaticity coordinate values obtained as described above into the following equation (1).

[0036]

(Equation 1) The relative spectral distribution of the light source is obtained by the following equation (2) for obtaining the spectral distribution of the perfect radiator, and the correction according to the type of the light source performed in step S4 is thereby performed. It will be.

[0037]

(Equation 2) In the embodiment configured as described above, the transmission condition of the neural network for determining the paint color is changed according to the contents of the selected database, and the determined paint color is also different. Each database is divided into regions, genders, ages, and so on. It is possible to make it.

When all of these databases are selected, the paint color is determined by the most quantified database on average, and the reproducibility of the determined color is increased.

It is also conceivable to use a personal database of the designer. In this case, a color determined based on the sensitivity unique to the designer can be used as a candidate color, or the candidate color can be determined based on the averaged sensitivity of a plurality of designers.

Since the present invention is configured as described above, it has the following effects.

According to the first aspect of the present invention, there is provided a color simulator capable of displaying a candidate color having a regional characteristic or a candidate color reflecting a user's preference only by inputting an instruction for selecting a sensibility term and a database. There is an effect that can be.

According to the second aspect, since the display color is changed according to the type of the light source, the color can be evaluated under various situations, and the effect that the simulation can be performed more finely can be obtained. is there.

[Brief description of the drawings]

FIG. 1 is a diagram showing a functional configuration of a main part of an embodiment of the present invention.

FIG. 2 is a diagram showing a functional configuration showing a process of generating and coloring a model in the embodiment of the present invention.

FIG. 3 is a flowchart illustrating a control operation of a control device when a display image is generated according to the present invention.

FIG. 4 is a diagram showing a configuration of a neural network for determining a display color.

FIGS. 5A and 5B are diagrams illustrating a method of converting a Munsell color into an RGB signal according to the present invention.

[Explanation of symbols]

 Reference Signs List 101 Display color determination means 102 Paint color database 103 Sensitivity database group 104 Input means 105 Munsell color-R, G, B conversion means 106 Display system S201 Modeling processing S202 Rendering processing S203 Library making processing S204 Painting processing S1 to S4 Step

──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Kazuo Aoki 2-6-3 Otemachi, Chiyoda-ku, Tokyo Inside Nippon Steel Corporation (56) References JP-A-1-318173 (JP, A) JP-A-2-170281 (JP, A) JP-A-3-123274 (JP, A) JP-A-4-27832 (JP, A) Masahiro Yamamoto and four others, "Status of AI Basic Research", NEC Technical Report , NEC Cultural Center, Inc., May 15, 1992, Vol. 45, No. 4, p. 105-119 (58) Field surveyed (Int. Cl. ⁷ , DB name) G06F 17/50 G06T 1/00

Claims (2)

(57) [Claims] 1. A design evaluation color simulator for displaying candidate colors to be used for evaluation according to sensibility terms, wherein a large number of unspecified humans classified by region, gender, age, etc. are provided with a "color that feels fun." , A "sensible color", etc., based on the results of a survey, etc., a sentiment database group consisting of a plurality of databases in which sentiment terms and colors are associated with each other; An input means for performing, a neural network including an input layer corresponding to a kansei term, and an output layer corresponding to a plurality of candidate colors, and determining a candidate color to be used for evaluation according to the input content to the input means. Display color determining means, wherein the display color determining means is configured to control the neural network in accordance with the contents stored in the database selected by the input to the input means. Performs weighting network to configure a network, color simulator design evaluation and determining the candidate color used for evaluation using the neural network. 2. The design evaluation color simulator according to claim 1, wherein the candidate color determined by the display color determination means is converted in accordance with an instruction input indicating the type of light source input to the input means. A color simulator for design evaluation, comprising: