JPH10145619A - Color simulation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To simulate a color in response to an environment when an original image whose color separation value is known is observed. SOLUTION: A color information value of an input image obtained by picking up an object is converted into at least three major component vector coefficients corresponding to the color information value with a neural network whose learning is finished. The major component vector coefficient is multiplied with the stored major component vector to convert the product into a spectral reflectance with a weight depending on a human color discrimination capability, and the weighted spectral reflectance is divided by a stored weight coefficient to exclude the weight and the result is converted into a spectral reflectance of the object.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color simulation method, and more particularly, to a color simulation method for reproducing the color of a photographed object on an image output device in consideration of human visual characteristics.

[0002]

2. Description of the Related Art For making a copy of a multi-colored document or object, or for displaying a multi-colored document or object,
One of the methods for expressing colors is a color separation value. Generally, these color separation values are YMCK (cyan, magenta, yellow, black) of subtractive color mixture for printing, and RGB (red, green, blue) of additive color mixture for display. By using the color separation values, each color of the original and the object can be reproduced. In recent years, a so-called color design for examining colors to be created has been performed before actually creating a multi-colored original or a copy of an object. A color simulation apparatus that reproduces each color of a document or an object and displays it on a display device is known as a device used for examining colors to be created. In this color simulation device, the color separation values of a large number of originals or objects are converted into device values of a display device, for example, TV signals or digital RGB output values and displayed. As a result, it is possible to display a multi-color document or a copy of an object in a reproduced color.

However, the above-mentioned color separation values depend on the characteristics of an apparatus for specifying the color of a document or an object and the characteristics of an apparatus for converting the color separation values into device values. For example, if a light source or a colorimeter for irradiating a document or an object to specify a color is different, the color of the document or the object has a different color separation value.
Further, if the conversion characteristic to the device value is different, the color of the document or the object is reproduced in a different color. Therefore, conventionally, the color to be reproduced does not depend on the characteristics of the device for specifying the color,
Regardless of the device, in order to express a value representing the same color as the color of a document or the like, the color is once converted to a device-independent color information value (intermediate expression method). Such a color as a device-independent color information value such as CIE (Commission Int.
ernationale deI'Eclairage)
Alternatively, there are color values such as a uniform perceived color space and CIE L * a * b * values.

Using color values such as CIE XYZ values or CIE L * a * b * values as color information values, color separation values such as CMYK values representing the color of a document or an object are converted into color values. later,
Colors are reproduced by so-called device-independent color correction. As described above, by matching the color values independent of the device, it is possible to reproduce colors with stable accuracy even in different color simulations for a document of many colors (Japanese Patent Laid-Open No. 4-212267).

[0005]

However, CIE L *
The a * b * and CIE XYZ values are defined by the spectral distribution of the light source, the spectral reflectance and spectral transmittance of the target surface of the document or object, and the color matching function. The illumination light when observing the color of is CIEL * a * b *
If the illumination light is not the same as the predetermined illumination light used for deriving the value, there is a problem that the reproduced color does not match the color of the document or the like. As described above, in the conventional device-independent color simulation apparatus, the light source whose reproduced color appears to be the same as the color of the document or the object is limited only by the light source of the predetermined illumination light. There is a problem that colors can be reproduced only under an environment (under illumination). That is, illumination for observing the color of a document or object is generally under an incandescent light bulb, under a fluorescent light,
It is various, such as under sunlight. Therefore, in the conventional device-independent color simulation apparatus, it is difficult to always match the colors to be reproduced under such various light sources. For this reason, even when the color simulation apparatus displays a document to which a color separation value by YMCK for printing is given on a display device, for example, only a color image with a fixed light source can be observed, and the color under a different light source can be observed. It was not possible to simulate each color of the image.

When a subject is photographed using a photographing device such as a video camera or a digital camera, a signal corresponding to an original image output from the photographing device is a color of an illumination light source at the time of photographing the subject. The signal becomes colored depending on the characteristics. Therefore, when the subject is visually observed by displaying the original image on a monitor such as a television using a signal from the photographing device, the displayed original image is displayed in a color different from the color of the subject. For example, even if the same subject is photographed under standard sunlight in a room such as a fluorescent lamp and under sunlight, the subject is reproduced in different colors. The present invention is a color simulation method for reproducing the color of a subject in consideration of human visual characteristics when outputting an original image of the photographed subject to an output device for observation in consideration of the above fact.

[0007]

According to a first aspect of the present invention, a color information value of an input image obtained by photographing a subject is subjected to multivariate analysis of spectral reflectance weighted according to human color discrimination ability. And converting the characteristic parameter coefficients into the weighted spectral reflectance based on the at least three characteristic parameters obtained by the multivariate analysis. , And the weights are removed from the image data and converted into the spectral reflectance of the subject. According to a second aspect, in the color simulation method according to the first aspect, principal component analysis is used as multivariate analysis, and a principal component vector of the weighted spectral reflectance obtained by performing the principal component analysis is used as a feature parameter. , A development coefficient of the weighted spectral reflectance based on the principal component vector is used as the feature parameter coefficient.

According to a third aspect of the present invention, in the color simulation method according to the first aspect, the color information values of the input image obtained by the photographing are converted from the color information values to the at least three features corresponding to the color information values. The characteristic parameter coefficients are converted by a multi-layer feedforward neural network that has been learned to be converted into parameter coefficients. According to a fourth aspect, in the color simulation method according to the first aspect, the color information value of the input image obtained by photographing is obtained by combining the color information value and the at least three characteristic parameter coefficients corresponding to the color information value. Are converted into the characteristic parameter coefficients by a look-up table that stores the corresponding relations.

According to a fifth aspect of the present invention, in the color simulation method according to the first aspect, the human color discrimination is performed by multiplying a spectral reflectance as a color matching function or chromaticity coordinates or a visual cell characteristic as a weighting factor. It is characterized in that the spectral reflectance is weighted according to the performance. According to the color simulation method of the present invention, first, color information values of an input image obtained by photographing a subject are converted into at least three characteristic parameter coefficients corresponding to the color information values. That is, each color of the input image can be approximated by at least three characteristic parameters and characteristic parameter coefficients obtained by performing a multivariate analysis of spectral reflectances weighted according to human color discrimination ability in advance. In this way, by expressing with the feature parameters, the color can be expressed without depending on the inherent color characteristics depending on the photographing method.

As the multivariate analysis, a principal component analysis can be used, and a principal component vector of the weighted spectral reflectance obtained by the principal component analysis is used as a feature parameter. The expansion coefficient of the weighted spectral reflectance based on the principal component vector can be used as a feature parameter coefficient. The weighting of the spectral reflectance according to the human color discrimination ability is performed as described in claim 5, wherein the color matching function or the chromaticity coordinate or the photoreceptor characteristics is used as a weighting factor. Can be multiplied by In other words, color matching functions or chromaticity coordinates by the CIE 1931 colorimetric standard observer or CIE 1964 colorimetric auxiliary standard observer specified by the CIE, or the light receiving characteristics of photoreceptors are used as weighting factors according to human color discrimination ability. Weighting can be performed by multiplying the spectral reflectance.

In the first conversion, a multi-layer feedforward neural network learned to convert the color information value of the input image into at least three characteristic parameter coefficients corresponding to the color information value is used. Can be used. As is well known, the neural network outputs an output reflecting the learning regardless of any input from learning of a large number of input / output relations. Therefore, if a color information value is input to the first conversion, Outputting at least three feature parameter coefficients corresponding to the color information value; The at least three feature parameter coefficients are calculated based on the at least three feature parameters obtained by the multivariate analysis.
In the following conversion, the weighted spectral reflectance is converted. That is, a weighted spectral reflectance is obtained by multiplying each of at least three characteristic parameters by a characteristic parameter coefficient.

From the weighted spectral reflectance, the weight is removed by a third conversion to convert the spectral reflectance into the spectral reflectance of the subject. Since the converted spectral reflectance does not depend on the light source, each color of the input image can be specified.

[0013]

DESCRIPTION OF THE PREFERRED EMBODIMENTS First, the concept of the present invention will be described. The process by which a person recognizes the color of an object is to collect the reflected light from the object reflecting light from a light source of a predetermined color with the eyeball,
It is transmitted to the brain via various visual filters and visual cells of the eyeball. Here, a spectral reflectance distribution or a spectral transmittance distribution of the object (hereinafter, collectively referred to as a spectral distribution),
The spectral distribution of the light source and the human visual characteristics are important points in recognizing colors. For example, when the same object is observed under different light sources, a human determines that the color is different.

On the other hand, research has been conducted on a method of approximating the spectral distribution of color using principal component analysis (KL expansion), and it is known that the spectral distribution can be accurately approximated with a small number of principal component vectors. By approximating and reproducing the spectral distribution of an object using this method of reproducing colors using principal component analysis, the color appearance under the desired light source considering the spectral distribution of the light source and human visual characteristics is considered. Can be simulated. Hereinafter, an example of an embodiment of the present invention will be described in detail with reference to the drawings. The present embodiment assumes that an original image of a subject represented by color information values such as RGB obtained by photographing the subject with a digital input device, video, or the like is observed under various light sources. The present invention is applied to a color simulation device that converts color information values of a photographed subject into colors for observation under various light sources, simulates them on a monitor or a printer, and outputs the color information values in consideration of information and characteristics of an output device. Applied.

As shown in FIG. 2, a color simulation apparatus according to the present embodiment includes a simulation main body 10 including a microcomputer, an input device 24 for inputting color data of a photographed subject, and an image data input / output device.
26, keyboard 28 for inputting data and commands,
And a monitor 30. This input device 24
A digital camera, a video camera, and a still camera are used for photographing a subject such as a document or an object and inputting color data using the photographed subject as an input image. Input device
The data input from 24 includes digital data or analog data of each color of RGB independently or including the synchronization signal used for the display device for the computer, the synchronization signal is separated, and the TV signal and the video signal are There is a TV signal (for example, NTSC, PAL or the like as a signal standard) including a synchronization signal.

The simulation body 10 includes a CPU 12
, A ROM 14, a RAM 16, a memory 18 for storing color information of a light source and the like (details will be described later), and an input / output device (hereinafter referred to as I / O) for exchanging data and the like between the simulation main body and other devices. 20 and a bus 22 connected to these so that data and commands can be input and output. The ROM 14 stores a processing program described later. After the learning of the simulation body 10 described below is completed, only the input device 24 may be provided. That is, the image data input / output device 26
Is a device for directly inputting colors as numerical data, because the simulation device is used for conversion learning for reproducing and displaying irrespective of the color characteristics of the input device. This is a device in which the decomposition values are stored in, for example, external storage means and read from this external storage means.

FIG. 1 is a block diagram showing a schematic configuration of each function of the color simulation apparatus 50 according to the present embodiment. In the color simulation device 50 of the present embodiment, the input device 24
An original image based on color separation values such as RGB input from the computer is simulated into an image for observation under a desired light source designated by an operator and reproduced by the output device 30. The conversion from the color separation values to the output device values in the color simulation device 50 is performed by the first conversion unit 32 and the second conversion unit for each function.
34, a third converter 36, a fourth converter 38, and a fifth converter 40. The first conversion unit 32 is for converting the color separation values input from the input device 24 into principal component vector coefficients, and the second conversion unit 34 is input using the converted principal component vector coefficients. The third conversion unit 36 calculates the weighted spectral reflectance corresponding to the color separation value according to the human color discrimination ability, and removes the weight using a weight coefficient 44 stored in advance. And the fourth conversion unit 38
Is for obtaining a color value under the specified light source using the obtained spectral reflectance and the light source information 42 stored in advance, and the fifth conversion unit 40 outputs the obtained color value to the output device 30. It is for converting to a device value.

The first conversion section 32 has a conversion function constituted by a neural network described later,
It has a learning function to learn it. Next, details of each functional unit will be described. The first conversion unit 32 is a conversion system constructed in advance for conversion of color separation values as color information values, and converts color separation values such as RGB of an original image obtained by photographing a subject from the input device 24. Conversion into expansion coefficients of at least three principal component vectors. For this conversion, a neural network can be used. The first conversion unit 32 has neurons corresponding to the number of color separation values as input layers for inputting color separation values for each of the RGB colors, and converts coefficient values (principal component vector coefficients) through an intermediate layer. A neural network according to the number of feature parameter coefficients is provided as an output layer for outputting, and a neural network in which each neuron is connected by a synapse is learned by a learning process described below, and unknown from a desired color separation value. To obtain a system for calculating the principal component vector coefficients of

As shown in FIG. 3, the first conversion unit 32 includes a network 46 and a teacher unit 48. The network 46 receives the color separation values for each of the RGB colors and outputs the estimated principal component vector coefficients. The teacher signal TC corresponding to the input color separation value and the output signal OC corresponding to the output principal component vector coefficient are input to the teacher unit 48, and the teacher unit 48 converts the correction signal SC obtained from the difference and the like into the network 46. Output. That is, the color separation values are input to the network 46 for each of the RGB colors, and the estimated principal component vector coefficients are output. A known principal component vector coefficient corresponding to the input color information value is used as a teacher signal TC as a teacher unit 48.
The output signal OC corresponding to the output principal component vector coefficient is also input to the teacher unit 48. The teacher unit 48 transmits the corrected signal SC obtained from the difference between these input signals to a network.
Output to 46.

As an example of the neural network used in the first conversion unit 32, as shown in FIG.
Three units of R, G and B as color separation values (or
An input layer composed of Y, M, C, K and the number of components of each video signal component) I1, I2, I3, and a large number of units Mq (q
≧ 1), and output units U1, U2,.
-It consists of an output layer consisting of Un. Each unit of the intermediate layer is connected to a bias unit. The unit of the intermediate layer is constituted by a neural circuit element whose input / output relationship is represented by a sigmoid function, and the unit of the input layer and the unit of the output layer are constituted by neural circuit elements whose input / output relationship is linear. The unit of the output layer may be formed of a neural circuit element having an input relationship having a sigmoid characteristic similarly to the unit of the intermediate layer. The number of output units in the output layer is
Among the feature parameters obtained by multivariate analysis of the spectral reflectance distribution, the number of feature parameters used to actually represent the spectral reflectance, that is, the number of feature parameters corresponds to the principal component vector coefficients as described later.

In the first converter 32, when a device value such as an RGB value or a CMYK value is given as an input, an output Hj according to the following equation is output from a unit in the middle layer of the network.

[0022]

Where W (h) ij is the coupling coefficient (ie, weight) of the i-th unit of the input layer and the j-th unit of the intermediate layer
, Ii is the output of the ith unit of the input layer (ie, the color information value), and bj is the bias value supplied from the bias unit. Similarly, the output Ok of the output layer unit is also expressed by the following equation.

[0023]

## EQU2 ## where W (o) jk represents the coupling coefficient of the j-th intermediate unit and the k-th output unit, and Hj is
Is the output value of the j-th unit in the hidden layer defined by. Further, n, m, and u are the numbers of units in the input layer, the intermediate layer, and the output layer, respectively. Therefore, by inputting the color separation values to the units of the input layer, the values of at least three principal component vector coefficients are output as outputs Ok from the units of the output layer.

The characteristics of each unit in the input layer are such that the input is output as it is, and the coupling coefficient of the unit in the intermediate layer is between [0,1] defined by the following equation. This is a characteristic represented by a monotonically increasing sigmoid function.

[0025]

## EQU3 ## The unit of the output layer has a characteristic represented by a sigmoid function or a linear function. However, the output layer may have an input from the bias unit. Coupling coefficients of this neural network are learned in advance so as to minimize the error between the output of each unit of the output layer and the teacher signal, which is a characteristic parameter coefficient obtained by principal component analysis of the weighted spectral reflectance. Will be modified. That is, conversion from RGB values, CMYK values, and the like to principal component vector coefficients, which are characteristic parameter coefficients obtained by principal component analysis of the weighted spectral reflectance, is learned.

In the present embodiment, a feedforward connection type neural network is used as the neural network, and there are various methods for learning the network. For example, a back propagation algorithm (Runmelhert, DE and McClell
and, JL (Eds), "Parallel Distributied Processin
g, "Exploration in the Microstructure of Congnitio
n. Vol.1, 2, MIT Press, Cammbridge (1989)])
The steepest descent method can be used. Here, the principal component vector obtained by the principal component analysis is defined as a solution of an eigen equation defined by the following equation by applying the above-mentioned weighting to the spectral reflectances of various colors and using the covariance matrix.

[0027]

Where F is a covariance matrix defined by the following equation, l
Represents an eigenvalue, and v represents an eigenvector (principal component vector).

[0028]

## EQU5 ## The characteristic parameter coefficient is defined by the following equation using the principal component vector obtained from equation (4).

[0029]

Where Yi represents a feature parameter, and Ri
Represents an average vector of the weighted spectral reflectance. Next, details of the neural network learning process in the first conversion unit 32 will be further described with reference to FIG. First, as processing S1 before learning the neural network, the spectral reflectance distribution of a color chart with a known color separation value such as RGB or YMCK is measured. In the next process S2, the measured spectral reflectances of the plurality of color patches are weighted. Next processing
In S3, the weighted spectral reflectance is subjected to principal component analysis. Next, in process S4, at least three principal component vectors and principal component vector coefficients obtained by performing principal component analysis are derived. This principal component vector is defined as the solution of the eigen equation defined by Equation 4 as described above. The principal component vector coefficient is a principal component vector coefficient Y calculated by Expression 6. In the next process S5, at least three principal component vector coefficients of the color chart obtained in the above process are used as teacher signals, and color separation values such as RGB captured by the input device 25 are input to the neural network as input signals, respectively. Learning is performed in the next process S6. That is, learning is performed so as to minimize the square error between the output of the network and the principal component vector coefficient (feature parameter) teacher signal obtained by the principal component analysis. After the above processing is completed (processing S7) and the learning of the neural network is sufficiently performed, the structure and weight of the network are stored in the memory 18 in processing S8, and the construction of the conversion system is completed. In this way, by learning, the first conversion unit 32 allows the input device
When the color separation values such as RGB output by the shooting of 24 are input, at least three principal component vector coefficients corresponding to the color separation values output by the input device 24 are set to values independent of the color characteristics of the input device 24. Output as

Next, the second conversion unit 34 uses at least three principal component vector coefficients output from the first conversion unit 32 and at least three principal component vectors derived by the above-described principal component analysis. , And has a function of deriving the weighted spectral reflectance, and outputs the weighted spectral reflectance converted according to the following equation.

[0031]

## EQU7 ## Next, the third conversion unit 36 reads the weight coefficient stored in the memory 18 in advance, and divides the weighted spectral reflectance output from the second conversion unit 34 by the read weight coefficient. Thus, the weight is removed and the spectral reflectance is output. The fourth conversion unit 38 uses the spectral reflectance of the desired light source stored in advance in the memory 18 or the spectral information of the light source obtained by the measurement, and calculates the spectral reflectance from the third conversion unit 36 under the light source. It has a function to convert to color values,
The color value converted according to the following equation is output. Note that as the color values, CIE XYZ, CIE L * a * b *, CIE L * u * v *, etc. can be used. In the present embodiment, the tristimulus values of CIE XYZ are used.

[0032]

Where S (l): light source information (spectral distribution of the light source) R (l): spectral reflectance ____x, y, z output from the third conversion unit: standard visual function fifth conversion The unit 40 is for converting the color value signal from the fourth conversion unit 38 into an output value suitable for outputting an image using the characteristic value of the output device. In the present embodiment,
A matrix operation for obtaining RGB values required for display on a monitor, or a neural network for obtaining CMYK values required for output by a color printer can be used.

Next, the operation of the color simulation apparatus 50 according to this embodiment will be further described with reference to the flowchart of FIG. After the learning of the first conversion unit is completed as described above, when the power of the color simulation apparatus 50 is turned on or an instruction to start the simulation is issued from the keyboard, the process proceeds to step 100 in FIG. The color separation values of the original image of the document or object to be scanned are read. In the next step 101, the first of the neural network trained as described above
In the conversion unit 32, the input color separation value is at least 3
It is converted into two principal component vector coefficients, and the converted principal component vector coefficients are output to the second conversion unit 34. Then, step
In step 102, as described above, the previously derived principal component vector (stored in the memory 18) is read, and at least three principal component vector coefficients output from the first transform unit 32 are read into at least three principal component vectors. The weighted spectral reflectance is output to the third conversion unit 36 by multiplying the two principal component vectors. In step 103, the weight coefficient previously stored in the memory 18 is read, and the second
The weighted spectral reflectance output from the converting unit 34 is divided by the read weight coefficient to remove the weight, and is converted into a spectral reflectance corresponding to the original image. In the next step 104, by reading the operator's keyboard instructions,
Upon receiving an instruction under which light source the original image input by the input device 24 is simulated, the light source information of the specified light source is read from the memory 18. In the next step 105,
The fourth conversion unit 38 uses the spectral reflectance of the original image output from the third conversion unit 36 and the read spectral information of the light source or the spectral information of the light source obtained by measurement to obtain a desired light source. Convert to color values. In the next step 106, the fifth converter 40 outputs a color value under a desired light source to an output device.
In order to output an image to 30, an output value is converted using the characteristic value of the output device 30. In the next step 107, the fifth conversion unit
According to the output value from 40, the image is output in a color at the time of observing the original image under a desired light source.

As described above, in the color simulation apparatus according to the present embodiment, the first conversion unit obtains the principal component vector coefficient of the weighted spectral reflectance determined from the color separation value of the original image, and calculates the principal component vector coefficient. The weighted spectral reflectance is obtained by the second conversion unit using the above, the weight is removed by the third conversion unit,
The color value is obtained from the light source information of the light source to be irradiated by the fourth conversion unit, and the fifth conversion unit performs conversion according to the characteristics of the output device, and then outputs the original image. It can be output as an original image equivalent to the case of observation under various light sources. Therefore, by using the present color simulation device, it is possible to reproduce the original image captured by the camera with the output device regardless of the light source for observing the original image.

The color simulation apparatus according to the present embodiment simulates an original image assuming a case where a destination guide board, an emergency guide board, or the like is observed in a special environment where the color of an illumination light source such as a dark room is special. It is suitable for. In the above embodiment, the simulation apparatus includes a neural network in the main body. However, the neural network may be used as a separate and independent configuration. Further, in the above embodiment, the case where the input device is directly connected to the simulation device main body has been described.
Simulate the reproduction signal by connecting a device that captures the subject with a video camera or the like, performs magnetic recording or optical recording with a video signal (TV signal) or digital signal, and reproduces the magnetic recording medium or optical recording medium that is the recording medium. You may make it input into an apparatus main body.

[0036]

As described above, according to the present invention, at least the color information value of the input original image can be obtained by multivariate analysis of the spectral reflectance weighted according to the color discrimination ability of a human. It is possible to obtain the spectral reflectance of the original image by converting it into three characteristic parameter coefficients and removing the weight from the weighted spectral reflectance obtained using the characteristic parameter and the characteristic parameter coefficient.
By converting to a color value based on the spectral information of the specified light source and outputting it, the subject photographed for inputting the original image with the input device can be output in the color observed when observed under the specified light source. This has the effect.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating an outline of a functional configuration of a color simulation apparatus.

FIG. 2 is a block diagram illustrating a schematic configuration of a color simulation apparatus according to the embodiment;

FIG. 3 is a conceptual image diagram for explaining the operation of the neural network.

FIG. 4 is a conceptual image diagram for explaining a configuration of a neural network.

FIG. 5 is a flowchart showing a processing procedure when constructing a conversion system of a neural network as a first conversion unit.

FIG. 6 is a flowchart illustrating a flow of an operation of the color simulation apparatus.

[Explanation of symbols]

 10 Simulation main unit 12 CPU 14 ROM 16 RAM 18 Storage memory 20 Data input / output device 22 Bus 24 Original image input device 26 Image data input / output device 28 Keyboard 30 Original image output device 32 First conversion unit 34 Second conversion unit 36 No. 3 conversion unit 38 4th conversion unit 40 5th conversion unit 42 Light source information 44 Weight coefficient 46 Neural network 48 Teacher unit 50 Color simulation device

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁶ Identification code FI H04N 9/74 G06F 15/66 310 H04N 1/46 Z

Claims (5)

[Claims] 1. A color simulation method for reproducing a color of a photographed object on an image output device in consideration of human visual characteristics, wherein a color information value of an input image obtained by photographing the object is determined by a human. Is converted into at least three characteristic parameter coefficients obtained by performing a multivariate analysis on the spectral reflectance weighted according to the color discrimination ability of the image processing apparatus, and the converted at least three characteristic parameter coefficients are obtained by a multivariate analysis. A color simulation method comprising: converting to a weighted spectral reflectance based on at least three characteristic parameters obtained; removing weights from the converted weighted spectral reflectance; and converting the weighted spectral reflectance into a spectral reflectance of a subject. 2. A principal component analysis is used as a multivariate analysis, and a principal component vector of a weighted spectral reflectance obtained by the principal component analysis is used as a feature parameter, and a development coefficient of the weighted spectral reflectance by the principal component vector is calculated. 2. The color simulation method according to claim 1, wherein the parameter is a characteristic parameter coefficient. 3. A multi-layer feedforward neural network that has been learned to convert a color information value of an input image obtained by photographing a subject from the color information value to the at least three characteristic parameter coefficients corresponding to the color information value. 2. The color simulation method according to claim 1, wherein the parameter is converted into a characteristic parameter coefficient by a network. 4. A color information value of an input image obtained by photographing a subject, the color information value and at least three color information values corresponding to the color information value.
2. The color simulation method according to claim 1, wherein the conversion into the characteristic parameter coefficients is performed by a look-up table storing the correspondence between the characteristic parameter coefficients. 5. A method of weighting a spectral reflectance according to a human color discrimination ability by multiplying a spectral reflectance by a color matching function or a chromaticity coordinate or a photoreceptor characteristic as a weighting factor. The color simulation method according to claim 1, wherein