JP2002271640A - Image processor and image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing method and an image processor that can highly precisely convert an input image into a color space, independently of the processor or the processor and lighting. SOLUTION: In this method, an image signal with a color chart received in advance is used to decide a set of colors converted by the same conversion parameter in the case of converting the image signal inputted from an image input device, and the conversion parameter is calculated by heavier-weighting colors in the set of colors and a lighter-weighting colors at the outside of the set of colors, to discriminate as to which color set each pixel of the inputted image signal belongs at conversion, thereby conversion processing is conducted by using the conversion parameter of the discriminated set.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an image processing apparatus and an image processing method for converting color image data input from an image input apparatus into a color space independent of the apparatus or the apparatus and the illumination with high accuracy. is there.

[0002]

2. Description of the Related Art In recent years, various devices, such as scanners, digital cameras, printers, and displays, have been used as devices for processing color images. According to a technique for exchanging image data between these devices, color image data input from an input device is once converted into an independent color space independent of the device, and then converted into color image data to be output to an output device. There is something.

As described above, if the conversion between the signal of the image input device and the color space independent of the device is established, the data can be transferred to any image output device. There is no need to determine the number of color conversion processes.

If the color image data input from the image input device is converted into an independent color space not depending on the device but also on the illumination, an image under illumination different from the illumination at the time of image input can be output to the output device. You can also output from.

[0005] As an independent color space independent of the device,
XYZ tristimulus values specified by the International Standards Institute CIE, and L ^* a ^* b ^* color system, L ^* u ^* v ^* color system, or it is common to use a color appearance model such as CAM97s, Since the attribute values of the color appearance model are calculated from the XYZ tristimulus values, if the XYZ tristimulus values can be estimated from the signals of the image input / output device, the above color conversion becomes possible.

As a color space that does not depend on a device and illumination, it is general to use the spectral reflectance of an object. XYZ tristimulus values can be calculated by integrating desired illumination on the spectral reflectance.

Estimating the XYZ tristimulus values or the spectral reflectance of an object from the color space of each image input / output device in this way is called characterization.

The present invention relates to the characterization of an image input device such as a digital camera, a multispectral camera, and a scanner.

Conventional characterization techniques for image input devices include, for example, a colorimetric method for skin and a method for estimating a spectral reflection spectrum described in Japanese Patent Application Laid-Open No. H7-174631, and a color reproduction method described in Japanese Patent Application Laid-Open No. H11-89552. There is a color simulation device described in JP-A-9-233490.

Japanese Patent Application Laid-Open No. 7-174631 discloses a method of estimating a spectral reflection spectrum of skin from an image input from an image input device.

The procedure will be described with reference to FIG.

First, the image data RGB input in step 1101 is converted into a signal linear in luminance by a quadratic function. The quadratic function described in the above publication is shown in (Equation 1). (Equation 1) is determined so that the XYZ tristimulus values of the 9-tone color chip of achromatic color are measured, and the luminance is linear with the Y value.

[0013]

(Equation 1) Next, in step 1102, XYZ tristimulus values are calculated from the luminance linear signal by a multiple regression matrix using at least the second-order terms. Finally, in step 1103, the spectral reflectance is estimated from the XYZ tristimulus values.

The multiple regression matrix in step 1102 needs to be determined in advance. In order to determine the multiple regression matrix, first, as a specific object, the skin is photographed with an image input device, the image data is obtained, and the colorimeter is also used to measure the color to obtain XYZ tristimulus values. Then, convert the image data to X
A matrix M to be converted to YZ tristimulus values is estimated by the conversion.
The determination is made so that the error between the XYZ tristimulus values and the XYZ tristimulus values measured by the colorimeter is minimized. Determining the estimation matrix so that the error between the predicted value and the actually measured value is minimized is called multiple regression analysis, and the estimation matrix determined in this way is called multiple regression matrix. The multiple regression matrix is represented by (Equation 2) where T is the XYZ tristimulus value vector and I is the image data vector. In (Equation 1), R _TI represents a correlation matrix between T and I.

[0015]

(Equation 2) In addition, the dimension of the spectral reflectance in the step 1103 is considerably large, for example, 31 dimensions even if the range of visible light from 400 nm to 700 nm is sampled every 10 nm, and it is difficult to estimate. Is also represented by a low-order m-order basis. Since the cumulative contribution of the spectral reflectance of the skin as the object to the third principal component is 99.5%, m
Since = 3 is sufficient, the base coefficient can be uniquely obtained from the XYZ tristimulus values.

As described above, in the conventional characterization method, the object is limited to the skin, and a matrix for estimating the XYZ tristimulus values from the image data is obtained by multiple regression analysis of the image data of the skin and the actually measured XYZ tristimulus values. I have decided. Therefore, with this matrix, the XYZ tristimulus values of the skin can be estimated with high accuracy, but the XYZ tristimulus values of the skin other than the skin have an extremely large error.

In the color reproduction apparatus described in JP-A-11-89552, a matrix for obtaining XYZ tristimulus values from image data is derived as follows.

First, the tristimulus value T and the image data vector I
Can be expressed as (Equation 3). In equation (3), E _O lighting matrix during observation, X is the matrix of the color matching function and a lateral vector, f is the spectral reflectance, E _m is the illumination matrix at the time of photographing, S is an image input device spectral sensitivity This is a matrix as a horizontal vector. By substituting (Equation 3) into the multiple regression matrix (Equation 2), (Equation 4) is obtained. In this (Equation 4), R _ff is
It is a correlation matrix of the spectral reflectance of an object. A matrix for predicting XYZ tristimulus values from image data by calculating in advance the correlation matrix of the spectral reflectance of an object that is a main component of the input image and substituting it into R _ff in (Equation 4) (Equation 4) can be obtained.

[0019]

(Equation 3)

[0020]

(Equation 4) As described above, in the above-described characterization method of the image input device, the object is limited, and the matrix for estimating the XYZ tristimulus values from the image data is determined using the correlation matrix of the spectral reflectance of the limited object. . Therefore, when the XYZ tristimulus values of the image data other than the limited object are estimated using this matrix, the error becomes extremely large.

The color simulation apparatus disclosed in Japanese Patent Application Laid-Open No. 9-233490 discloses an illumination simulation in which an image input from an image input device is converted into a color under a desired light source and then output on a display. I have.

The procedure will be described with reference to FIG. Similar to the method described in Japanese Patent Application Laid-Open No. 7-174631, the spectral reflectance is subjected to principal component analysis, and is expressed by an m-order basis lower than the 31st order. Then, in step 1201, a m-th basis coefficient is estimated from the input image data by a neural network. Next, in step 1202, the spectral reflectance is calculated from the estimated m-th order vector. The obtained spectral reflectance is multiplied by a desired light source vector to obtain XYZ tristimulus values, which are converted into display drive signals using the color characteristics of the display.

In the neural network, an appropriate spectral reflectance is estimated if the input data has a property similar to the learning data, but an error becomes extremely large if the input data is not similar to the learning data. Therefore, this conventional example using a neural network can be said to be a method that can be estimated only when the object is limited.

[0024]

In the above conventional method,
In each case, limiting the object to one, determining a matrix or neural network for estimating the spectral reflectance,
XYZ tristimulus values or spectral reflectances are estimated for all pixels in the input image using one matrix or neural network.

However, it is extremely difficult to approximate all color patches to a simple linear model as in the conventional example.
There is a problem that the error becomes considerably large depending on the color chart. On the other hand, when a more complicated nonlinear model is used, that is, when the higher order terms of the input image are used, or when the number of neurons or layers is increased by a neural network, overfitting occurs and the color chart has a small error. However, colors other than the color chart are converted into extremely unique colors.

The present invention has been made in view of the above points, and has as its object to provide an image processing method for converting color image data input from an input device into a color space that does not depend on the device and illumination with high accuracy. And

[0027]

According to the present invention, when an image signal input from an image input device is converted, colors which can be converted with the same conversion parameters by using an image signal of an input color chart are set in advance. A color set is determined, and its conversion parameters are calculated by weighting the colors in the color set with a large weight and the colors outside the set with a small weight, and to which color set each pixel of the input image signal belongs at the time of conversion. Is determined, and the conversion processing is performed using the conversion parameters of the determined set.

In this manner, the color of the color chart is particularly accurate, and the color data independent of the apparatus or the apparatus and illumination can be estimated from the input image data with high accuracy even for colors other than the color chart.

[0029]

According to the first aspect of the present invention, when an image signal input from an image input device is converted, the same conversion parameter is used in advance by using the input image signal of the color chart. A color set determining means for determining a color set to be converted by: a parameter determining means for determining a conversion parameter for each color set in advance; a color set determining means for determining a color set to which the image signal belongs; And a conversion processing unit for converting the color of the pixel of interest using the conversion parameter of the color set determined by the unit. The color conversion can be performed with high accuracy on the image signal input from the image input device. It has the action of:

According to a second aspect of the present invention, when an image signal input from an image input device is converted, a color set to be converted with the same conversion parameters using an input color patch image signal in advance is used. Image processing characterized in that the conversion parameters are determined, and at the time of conversion, each color of the input image signal belongs to which color set, and conversion processing is performed using the conversion parameters of the determined set. The method has an effect that an image signal input from an image input device can be color-converted with high accuracy.

According to a third aspect of the present invention, in the conversion processing, the input image signal is converted into color data independent of an image input device or a device and illumination. The method has an effect that an image signal input from an image input device can be converted into high-precision device or color data independent of device and illumination.

According to a fourth aspect of the present invention, the determination of the color set is performed by removing color patches whose statistical properties do not match among the color patches. This is a processing method, and has the effect that all color chips and image signals input from the image input device can be color-converted with high accuracy.

According to a fifth aspect of the present invention, there is provided the image processing method according to the fourth aspect, wherein the statistical property is a normal distribution, wherein all the colors input from the image input device are used. This has the effect that the vote and the image signal can be color-converted with high precision.

According to a sixth aspect of the present invention, in the determination of the color set, first, a conversion parameter is calculated by using all of the color patches to obtain a conversion error, and the color patch from which the large conversion error is removed is determined. 6. The image processing method according to claim 2, wherein one color set is used, and each color chart having a large conversion error is used as a color set.
This has the effect that all color patches and image signals input from the image input device can be color-converted with high accuracy.

According to a seventh aspect of the present invention, in the determination of the color set, each of the color charts that are important in the conversion process is set as a color set. This is an image processing method, and has the effect that all color chips and image signals input from the image input device can be color-converted with high accuracy.

According to an eighth aspect of the present invention, there is provided the image processing method according to any one of the second to fifth aspects, wherein the color set is determined by a user.
This has the effect that all color patches and image signals input from the image input device can be color-converted with high accuracy.

According to a ninth aspect of the present invention, the conversion parameter for each color set is determined by calculating a statistical property of a color chart belonging to each color set. The image processing method described in (1) above has an effect that an image signal input from an image input device can be color-converted with high accuracy.

According to a tenth aspect of the present invention, the conversion parameter for each color set is determined by calculating a statistical property by weighting a color chart belonging to each color set to a large weight and a color chart not belonging to a color set to a small weight. An image processing method according to any one of claims 2 to 8, wherein an image signal input from an image input device can be color-converted with high accuracy.

According to an eleventh aspect of the present invention, in the conversion processing, a reliability belonging to each color set is calculated, and a value obtained by weighting the value converted by the conversion parameter of each color set with the reliability is used as a final conversion value. 11. The method according to claim 2, wherein
The image processing method according to any one of the above, which has an effect that the image signal input from the image input device can be color-converted with high accuracy.

According to a twelfth aspect of the present invention, the color space for determining the set to which the color of each pixel belongs is a color space of the image input device, a spectral reflectance color space, or a base reflectance color space of the spectral reflectance. The image processing method according to any one of claims 2 to 11, characterized in that the image processing method is an XYZ tristimulus value color space. Have.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

(Embodiment 1) In the present embodiment, a digital camera having RGB three-band outputs is assumed as the most general image input system, and an object which is data independent of a device and illumination from RGB image data. A method for estimating the spectral reflectance of the image will be described as an example. For the sake of simplicity, the case where the number of categories is 2 will be described.

FIG. 1 shows an input device according to the first embodiment.
FIG. 3 is a block diagram of an image processing device that estimates spectral reflectance image data from RGB image data.

In FIG. 1, reference numeral 101 denotes an image input device for acquiring image data of three RGB bands, reference numeral 102 denotes an input image, which is an image of a color chart at the time of pre-processing and an image of a conversion processing at the time of conversion. is there. Reference numeral 103 denotes an input image storage unit that stores the input image 102; 104, color patch image data which is image data input at the time of preprocessing; 105, based on color patch image data and spectral reflectance data of a known color patch. A color set determining unit that determines a color set including colors that can be converted with the same conversion parameter; 106, a color set parameter representing a range of the color set determined by the color set determining unit; 107, a color set parameter A color set storage unit for storing 106;
Reference numeral 8 denotes a conversion parameter determining unit for each color set which determines a conversion parameter for each color set. Reference numeral 109 denotes a conversion parameter for each color set determined by the conversion parameter determining unit 108 for each color set.
Reference numeral 10 denotes a conversion parameter storage unit that stores a conversion parameter 109 for each color set, 111 denotes target image data read out one pixel at a time from the conversion processing target image 102 stored in the input image storage unit 103, and 112 denotes target image data A color set determination unit 113 for determining a color set to which the 111 belongs;
Is a color set signal whose signal value is the color set number determined by the color set determination unit 112, and 114 is the spectral reflectance of the target image data 111 using the color set conversion parameter 109 specified by the color set signal 113. , A control signal instructing that the conversion process of the image data of interest has been completed, 117 an overall control unit for controlling the processing of each pixel of the input image 102, and 118 a spectral reflectance data An image recording unit for recording the image, 119 an image processing device, 12
0 is a spectral reflectance image output from the image processing device 119.

The operation of the image processing apparatus shown in FIG. 1 will be described.

The operation is divided into two types, a preprocessing for determining a color set from a color chart and an estimation processing for actually estimating a spectral reflectance image from an input image. First, the pre-processing will be described.

In the preprocessing, first, a color chart composed of a plurality of colors as shown in FIG.
The input image 102 is stored in the input image storage unit 103.

The input image 102 of the color chart stored in the input image storage unit 103 is passed to the color set determining unit 105 as color chart image data 104, and the color set determining unit 105 converts the color of the color chart into the same conversion parameter. A plurality of color sets are determined using the colors that can be estimated as one set. The color set determined by the color set determining unit 105 is stored in the color set storage unit 107 as a color set parameter 106. Also, the conversion parameter is passed to the conversion parameter determination unit 108 for each color set, and the conversion parameter is determined for each color set. Conversion parameter 1 for each determined color set
09 is stored in the conversion parameter storage unit 110. The above is the pre-processing.

The pre-processing is desirably performed when the image input device changes, when the illumination or environment of the subject changes, or when the type of the subject changes.

Next, the operation of estimating the spectral reflectance image from the input image will be described. An image of a subject is input from the image input device 101 and stored in the input image storage unit 103. Pixels are sequentially read out one by one from the input image storage unit 103 in accordance with an instruction from the overall control unit 117, and the target image data 11
It is passed to the color set determination unit 112 as 1. The color set determination unit 112 determines which color set the image data 111 of interest belongs to, and outputs the number value of the corresponding color set as a color set signal 113. The conversion parameter 109 is read from the conversion parameter storage unit 110 according to the signal value of the color set signal 113, and the spectral reflectance estimating unit 114 estimates spectral reflectance data 115 using the conversion parameter. The estimated spectral reflectance data 115 is recorded in the image recording unit 118. When the above-described process of estimating the target image data is completed, the control signal 116 is output to the overall control unit 117.
And the next image data of interest 111 is read from the input image storage unit 103. The above is repeated for all the pixels of the input image 102.

Next, details of the operations of the color set determining unit 105 and the color set-specific conversion parameter determining unit 108, which perform preprocessing for determining a color set, which is a feature of the present invention, will be described.

The color set determining unit 105 determines the range of colors that can be converted with the same conversion parameters, and there are a method of calculating using the color patch image data 104 and a method of specifying by the user. .

First, a method of calculating a color set using the image data 104 of the color chart will be described. The color chart is formed from the main constituent colors of the input image.

When the spectral reflectance estimation, which is the conversion process, is performed by multiple regression analysis, the conversion parameters are determined so that the mean square estimation error with the estimated value of the color chart used in the analysis is minimized. At this time, the color chart used for the analysis is a normal distribution,
It is known that the maximum likelihood estimation solution matches the estimation value by multiple regression analysis, that is, the least squares solution when the error of the model follows a normal distribution with a mean of 0.

Therefore, a plurality of colors forming a normal distribution are selected from the color charts as a color set, and a multiple regression distribution is performed with the same conversion parameter, thereby obtaining an optimal estimation result for each color of the color chart. We propose to find the matrix. An explanatory diagram of the color set is shown in FIG.

There are a plurality of methods for calculating a color set so that each color set forms a normal distribution. Here, the method is performed in the following procedure.

FIG. 4 shows a color set calculation procedure.

First, in step 401, a multiple regression analysis is performed using all color patches. The equation for estimating the spectral reflectance by multiple regression analysis is shown in (Equation 5).

[0059]

(Equation 5) In (Equation 5), x is a value obtained by discretizing the spectral reflectance, for example, when 380 nm to 780 nm are discretized every 10 nm (r1, r2, r
3, ..., r31) It is a 31-dimensional vector of ^t . x 'is an estimated value by multiple regression analysis. d is the RGB value of the image input device,
(r, g, b) ^t- dimensional vector or ^t- dimensional vector (r ² , g ² , b ² , rg, rb, gb, r, g, b, 1) extended to polynomial And Uppercase letters X and D mean a matrix in which the color chart vectors x and d used for analysis are arranged as columns. That is, in step 401, matrices X and D are created using all the color charts, and a matrix M as a conversion parameter is obtained. The spectral reflectance of the color chart used here is measured by a colorimeter in advance.

Next, in step 402, using (Equation 5),
The estimated value x ′ of the spectral reflectance of each color chart is calculated, and the square error with the actually measured value x is obtained.

In step 403, a color chart number whose square error is larger than a predetermined threshold is determined, and a color chart whose error is smaller than the threshold is determined as color set 1.

In step 404, among the color patches having a large error obtained in step 403, color patches having similar colors are grouped, and the grouped color patches are determined as different color sets. Name it.

In step 405, the color range indicated by the color set is calculated for the second and subsequent color sets. For example, the average color of the colors belonging to the color set is taken, and a range specified by, for example, the variable radius from the average color is set as a color range indicated by the color set. Variable radius
For example, 0.1 (when R, G, B are each represented by 0 to 1)
Is used, a range having a radius of 0.1 from the average color is the color range indicated by the color set.

The color set and the color range indicated by the color set obtained as described above are used as color set parameters 106 in the color set storage unit 1.
07. FIG. 5 is a conceptual diagram of the color range indicated by the color set.

It has been found from experiments that the distribution of the remaining color patches approaches a normal distribution by removing the color patches having a large error when actually converted by the conversion parameters. As described above, when the multiple regression analysis is performed on the color patches closer to the normal distribution, the estimation accuracy is improved for all the color patches. On the other hand, when an arbitrary color chart is removed, that is, when the distribution of the color chart is not considered, the number of color charts used for analysis is reduced, but color charts with higher accuracy and lower color charts are mixed. Is not preferred. This is a feature of the present invention.

A description will be given of another method of determining a color set by a user. Depending on the input image to be converted in the future, it may be possible to perform the estimation more efficiently if the user grasps the characteristics of the input image and specifies a color to be a color set. For example, when an input image having a lot of human skin is to be used, a lot of skin colors may be used for the color chart, and a color close to the skin color may be determined as one color set. FIG. 6 is an explanatory diagram in the case where a user designates a color set. On the display, an image of a color chart input from the image input device 101 is displayed. Each color chart is divided and assigned to a color set. At this time, the user designates a color chart of the color set 2 or later. For example, color set 2
Is assigned a color patch near the skin color, and a color set 3 is assigned a color patch near the green color of the trees. Color set 1 is color set 2,
The rest of the color chart specified in 3 is automatically registered.

This concludes the description of the method by which the user specifies a color set.

Next, the conversion parameter determination unit 10 for each color set
8 will be described in detail.

The conversion parameter refers to a matrix M in (Equation 5) when spectral reflectance estimation is performed by multiple regression analysis. First,
The conversion parameter of the color set 1, that is, the matrix M is given by (Equation 5)
Can be obtained by

Similar to the color set 1, the conversion parameters of the second and subsequent color sets may be obtained by forming matrices X and D using only the color charts belonging to the color set. However, such a method is often difficult for the color set 2 and subsequent color sets.
The reasons are the following two points.

(1) The number of elements in a color set is often small. In order to obtain the matrix M without any constraint, at least the same number of data as the number of dimensions of the vector d is required.

(2) If the color patches are sufficiently varied to reproduce the colors in the input image, only the color patches belonging to the color set alone may be used, but if not, the color patches belong to only the color set. If the conversion parameter is obtained only from the color chart, the color shifted from the color chart cannot be estimated well.

Therefore, in the present invention, in the second and subsequent color sets, color patches outside the color set are used in addition to color patches in the color set, and the color patches in the color set are weighted more heavily than the color patches outside the color set. Then, it is proposed to obtain the transformation parameter matrix M. Since the calculation of the matrix M is determined so that the mean square error of the color chart is minimized, the error of the color chart in the color set is weighted more than others to determine the mean square error.

More specifically, the conversion parameters of the second and subsequent color sets are obtained by the procedure shown in FIG. The case of color set 2 will be described as an example.

In step 701, the weight W to be given to the color patches in the color set 2 is set, and the color patches in the color set 2 are weighted more than the color patches outside the color set 2, and multiple regression analysis is performed by (Expression 5). Matrix M
Ask for. The color chart outside the color set 2 used at this time is the color set 1
It belongs to the color chart. FIG. 8 is an explanatory diagram of the matrices X and D by weighting. Thus, the data of the color chart in the color set 2 may be increased by the weight W.

Next, in step 702, the error between the estimated value of the spectral reflectance obtained by using the matrix M obtained in step 701 and the actually measured value is determined using the color set 2 and the color chart outside the color set 2.

If the error of the color chart in the color set 2 is larger than the predetermined threshold value in the step 703, the color chart whose error is larger than the predetermined threshold value among the color charts outside the color set 2 is determined. Is removed, and the matrix M is obtained again by Equation 1. Otherwise, the process is terminated, and the matrix M obtained at this time is determined as the conversion parameter of the color set 2.

In step 704, the matrix M obtained in step 702
Is used to determine the error between the estimated value of the spectral reflectance and the actually measured value using the color set 2 and the color chart outside the color set 2.

If the error of the color chart in the color set 2 is larger than the predetermined threshold value in the step 705, the value of the weight W is increased and the process returns to the step 701 again. Otherwise, the process is terminated, and the matrix M obtained at this time is determined as the conversion parameter of the color set 2.

The obtained conversion parameter of each color set, that is, the matrix M is stored in the conversion parameter storage unit 110.

As described above, the process is repeated until the error of the color chart in the color set 2 becomes smaller than the predetermined threshold.

The use of the conversion parameters obtained by the above-described color set determination and conversion parameter determination methods has an effect that an estimated value of an error smaller than a predetermined threshold can be always obtained for a color chart color. is there.

Next, the detailed operations of the color set determining unit 112 and the spectral reflectance estimating unit 114 which perform the actual conversion processing will be described.

The image data of interest 111 is stored in the color set determination unit 11.
In 2, the color set belonging to the color set range stored in the color set storage unit 107 is determined. Since the color set range is determined by the center color of the color set and its radius range radius, for example, if the center color of color set 2 is (R2, G2, B2),
When Expression 6 is satisfied, the image data of interest is determined to belong to the color set 2.

[0085]

(Equation 6) The set is determined from color sets 2 to n (n is the total number of color sets)
If it does not belong to any of the second and subsequent color sets, it is determined that it belongs to color set 1.

The number of the determined color set is the color set signal 11
3 is passed to the spectral reflectance estimating unit 114.

Although the above-described determination of the color set is performed in the RGB color space, the determination can be performed by performing the range of each color set in another color space. For example, it can be represented by the center color and radius of the spectral reflectance of a color chart belonging to a color set. Alternatively, it is possible to reduce the number of dimensions by expressing the spectral reflectance as a basis, and to express the spectral reflectance by the center color and radius of the basis coefficient of the spectral reflectance of the color chart belonging to the color set. Alternatively, it can be represented by the center color and the radius of the XYZ tristimulus values of the color chart belonging to the color set.

In these cases, a temporary spectral reflectance, a basis coefficient of the spectral reflectance, or an XYZ tristimulus value is calculated once from RGB using a matrix M calculated using the entire color chart. If the color set is determined from the provisional values, the same operation as in the case of the RGB space is sufficient.

Next, the spectral reflectance estimating unit 114 reads the matrix M based on the color set signal 113 from the conversion parameter storage unit 110, and estimates the spectral reflectance by (Equation 5).
FIG. 9 shows an explanatory diagram.

In the above description, the spectral reflectance is estimated by multiple regression analysis. However, a neural network or a penalty function to which some constraint condition is added is created, and the minimum value of the penalty function value is recursively searched. The method can also be performed. The conversion parameters differ depending on which estimation method is used.

FIG. 10 shows a system application example of the image processing apparatus of FIG. In FIG. 10, reference numeral 119 denotes an image processing apparatus disclosed in the present embodiment, 1001 denotes a display, and 1002 denotes a printer. An image of a subject is input by the image input device 101, and a spectral reflectance image 120 is output by processing in the image processing device 119. The spectral reflectance image 120 is passed to the display 1001 or the printer 1002, and is converted into a signal of each device.
1 or output from the printer 1002. The conversion of the signal from the spectral reflectance image 120 to each device is performed on the display 1001 or the printer 100.
2, or may be performed by a separate processing unit such as a personal computer and then transferred to the display 1001 or the printer 1002. Further, all the processing of the image processing apparatus 119 may be performed by the CPU inside the image input apparatus 101.

In the system shown in FIG. 10, the input image is converted once into color data that does not depend on the device or the device and the illumination, and then converted into the signal of each output device. Colors can be output to displays and printers.

As in this application example, when the obtained spectral reflectance image 120 is output to a display or a printer and observed by a human, a false contour may occur at the boundary between the regions of each category. This is because the color set was determined for each pixel and the spectral reflectance was estimated by a different method.

To solve such a problem, the following method may be adopted.

If it is determined that each pixel belongs to a color set other than the color set 1, the reliability of the color set
The result weighted by y may be used as the final spectral reflectance.

As the reliability, for example, a value obtained by dividing the distance from the center of the color set by the radius of the color set Radius may be used.

In the above description of the embodiment, the color data of the conversion destination is the spectral reflectance, but the XYZ tristimulus values may be directly estimated from the input image instead of the spectral reflectance. In this case, the spectral reflectance in the above description is used as it is.
What should be read as XYZ tristimulus values.

In the present embodiment, the description has been given of the image input section of the three RGB bands.
The present method can be similarly applied to not only B but also to a larger number. Further, the image input unit may be not only a digital camera but also a scanner or digitized data of an analog output, or the present process may be applied to each image of a moving image.

Further, according to the present invention, the operation of the image processing apparatus shown in FIG. 1 is stored in a CDROM, and a program stored in the CDROM is downloaded to a RAM on a PC.
This is to make the CPU above perform the processing of the color estimating means. Alternatively, it is stored in a ROM in the image input device, and causes the CPU in the image input device to perform the processing of the means.
In this case, the image data output from the image input device is not a color space display unique to the input device, but is image data in a color space independent of the device or the device and the illumination. This eliminates the need to install the color estimation means in the computer, and has the advantage that ordinary users who are not familiar with computers and color conversion can easily handle it. However, if the RGB image data of the input device can also be obtained by changing the mode, consistency with the conventional device can be obtained.

As described above, according to the present embodiment, when converting an image signal input from an image input device, the image signal of the input color chart can be converted in advance using the same conversion parameters. A color set is determined using the colors as a set, and its conversion parameters are calculated by weighting the colors in the color set with a large weight and the colors outside the set with a small weight, and determining which color of each pixel of the input image signal at the time of conversion. By determining whether a color belongs to a set, and performing a conversion process using the conversion parameters of the determined set, the color as a color patch is particularly accurate, and the colors other than the color patch are also accurately determined from the input image data by the device or the device and the device. Color data independent of illumination can be estimated.

[0101]

As described above, according to the present invention, there is provided an image processing method and an image processing method for converting color image data input from an image input device to a color space which does not depend on the device or the device and the illumination with high accuracy. A processing device can be provided.

[Brief description of the drawings]

FIG. 1 is a block diagram of an image processing apparatus according to a first embodiment of the present invention;

FIG. 2 is an explanatory diagram of a color chart according to the first embodiment;

FIG. 3 is an explanatory diagram of a color set according to the first embodiment;

FIG. 4 is a diagram showing an operation procedure of a color set determining unit 105 according to the first embodiment.

FIG. 5 is a conceptual diagram of a color set range according to the first embodiment;

FIG. 6 is an explanatory diagram of a color set user designation according to the first embodiment;

FIG. 7 is a diagram showing an operation procedure of a conversion parameter determining unit for each color set according to the first embodiment;

FIG. 8 is an explanatory diagram of a weighted multiple regression analysis according to the first embodiment;

FIG. 9 is a diagram illustrating a spectral reflectance estimating unit 114 according to the first embodiment;
Illustration of

FIG. 10 is a system application example diagram of the image processing apparatus according to the first embodiment;

FIG. 11 is a diagram showing an image processing procedure in Conventional Example 1.

FIG. 12 is a diagram showing an image processing procedure in Conventional Example 2;

[Explanation of symbols]

 Reference Signs List 101 Image input device 102 Input image 103 Input image storage unit 104 Color chart image data 105 Color set determination unit 106 Color set parameter 107 Color set storage unit 108 Each color set conversion parameter determination unit 109 Conversion parameter 110 Conversion parameter storage unit 111 Image of interest Data 112 Color set determination unit 113 Color set signal 114 Spectral reflectance estimation unit 115 Spectral reflectance data 116 Control signal 117 Overall control unit 118 Image recording unit 119 Image processing device 120 Spectral reflectance image

 ──────────────────────────────────────────────────続 き Continued on the front page F-term (reference) HB01 HB05 HB11 LA02 LB02 MA10 MA11 NA03 NA15 NA29 PA03 PA05

Claims (12)

[Claims] 1. A color set determining means for determining a color set to be converted with the same conversion parameter in advance by using an image signal of an input color chart when converting an image signal input from an image input device; Attention is paid to using a parameter determining means for determining a conversion parameter for each color set in advance, a color set determining means for determining a color set to which the image signal belongs, and a color set conversion parameter determined by the color set determining means. An image processing apparatus comprising: a conversion processing unit configured to convert a color of a pixel. 2. When converting an image signal input from an image input device, a color set to be converted with the same conversion parameter and its conversion parameter are determined in advance by using the input image signal of the color chart. An image processing method comprising: determining which color set each pixel of an input image signal belongs to at the time of conversion; and performing conversion processing using conversion parameters of the determined set. 3. The image processing method according to claim 2, wherein said conversion process converts the input image signal into color data independent of an image input device or device and illumination. 4. The image processing method according to claim 2, wherein the determination of the color set is performed by removing color patches whose statistical properties do not match among the color patches. 5. The image processing method according to claim 4, wherein said statistical property is a normal distribution. 6. The color set is determined by first calculating a conversion parameter using all color patches to determine a conversion error, and removing a color patch having a large conversion error into one color set. 6. The image processing method according to claim 2, wherein each color chart having a large conversion error is set as a color set. 7. The image processing method according to claim 2, wherein in the determination of the color set, each color chart important in the conversion process is set as a color set. 8. The image processing method according to claim 2, wherein the color set is determined by a user. 9. The image processing method according to claim 2, wherein the conversion parameter for each color set is determined by calculating a statistical property of a color chart belonging to each color set. 10. A conversion parameter for each color set, wherein color patches belonging to each color set are weighted heavily, and color patches not belonging to a color set are weighted small to calculate and determine a statistical property. Item 9. The image processing method according to any one of Items 2 to 8. 11. The conversion process according to claim 1, wherein a reliability value belonging to each color set is calculated, and a value obtained by weighting the value converted by the conversion parameter of each color set with the reliability is used as a final conversion value. Item 11. The image processing method according to any one of Items 2 to 10. 12. The color space for determining the set to which the color of each pixel belongs is a color space of an image input device, a spectral reflectance color space, a base expression color space of spectral reflectance, an XYZ tristimulus value color. The image processing method according to claim 2, wherein the image processing method is a space.