JPH11232430A - Color changing method for object image - Google Patents

Abstract

PROBLEM TO BE SOLVED: To change the texture peculiar to an original color into the texture peculiar to a changed color when the object color of an original image is changed into another color. SOLUTION: A declination characteristic table of an original color is prepared (S11) to show a correspondence secured between the declination representing a deviation of the direction of line of sight from the direction of regular reflection about the reflection of the original color surface. Then the declination of each part of an object is determined, based on the chromaticity of an original image and by making a reference to the table (S12). Meanwhile, a declination characteristic table of a changed color is prepared (S13). Based on this table, the chromaticity of the changed color is determined from the declination that is determined in the step S12 (S14). Thus, both color and texture can be simultaneously changed via the processing since the state where the chromaticity changes according to the declination corresponds to the texture of each color. In regard to an area where the gradation is deficient, the lightness of the image having a changed color is corrected, based on the lightness of the original image (S17).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for changing the color of an object image, and more particularly, to a method for changing the color of an object at the same time as changing the texture unique to each color.

[0002]

2. Description of the Related Art In recent years, color preferences of users have tended to be diversified, and the importance of outer panel color designs for automobiles has been increasing. At the same time, with the advent of new paint materials, it has become possible to create skin textures of various textures. Under such a background, the business contents of the skin color design are becoming more sophisticated, and it has been proposed to use computer graphic technology to support the business of designing the skin color.
The use of an automobile image makes it possible to examine the color of the outer panel without preparing a vehicle actually coated with the outer panel color.

[0003] In the design of the skin color, work such as observing a car painted with a new color or comparing a plurality of cars painted with different colors is performed. In such an operation, a color change process of changing the color of the original vehicle image to another color is effectively used.

[0004]

The color change processing includes:
For example, [1] "Accurate Rendering technique on Color
imetric Conception "(A. Takagi, H. Takaoka, T. Oshim
a, Y.Ogata, ComputerGraphics (SIGGRAPH'90), Vol.2
4, No.4, pp.263-272 (1990)), [2] "A CAD System fo
r Color Design of a Car "(T.Oshima, S.Yuasa, K.Sa
kanosita, Y.Ogata, EUROGRAPHICS'92, Vol.11, No.3,
(1992)), [3] "Development of color CAD system" (Shinji Yuasa, Kenichi Sakanoshita, Yoshinori Ogata, Tetsuya Oshima, Mikiharu Katai, Toru Ibaraki, Naoyuki Kobayashi, Eighth NICOGRAPH Contest Papers, pp.300- 309, Japan Computer Graphics Association
(1992)).

In these prior arts [1], [2], and [3], the measured values of the deflected spectral reflectance of the skin color and the shape CAD data (3D
A system for reproducing and displaying the appearance of a vehicle by a computer rendering method based on (CAD data) has been reported. However, when CAD data of a shape does not exist, these systems cannot be applied. There is also a disadvantage that it takes a lot of trouble and time to create CAD data of a shape.

If a general-purpose photo retouching tool is used,
Even without the shape CAD data, the color of the outer panel can be changed.
With this tool, it is possible to take a picture of the car to be designed, capture this picture into a computer and change the color of the car.

However, when using a general-purpose tool,
There is a disadvantage that color adjustment by visual observation is required, and the work requires much labor and time. Further, as described below, the general-purpose tool has a problem that the texture cannot be changed in accordance with the color change.

[0008] Here, "texture" means a state in which the color changes in each part of the body. Even if one outer panel color is painted on the entire body surface, the color observed by a human changes in each part of the body. If the skin color is different, the observed color change, that is, the “texture” is also different. For example, on a painted surface of a solid color, the highlight portion is narrow, and the brightness changes sharply at the edge of the highlight. On the other hand, on the painted surface of the metallic color, the highlight portion is wide, and the brightness changes relatively gradually at the edge of the highlight. Thus, each skin color has a unique texture.

In processing using a general-purpose tool, hue, saturation, and lightness are uniformly changed. At this time, lightness information and its change information are maintained. Therefore, it is difficult to change the texture when changing the color. In particular, when the color of the outer image is changed to a color of the outer plate having a texture greatly different from that of the original image, a good color-changed image cannot be obtained. This is because an unnatural image having no texture corresponding to the color after the change is generated.

In the color image control method described in Japanese Patent Application Laid-Open No. 5-40833, the color of an input image from a scanner or the like is changed, and the texture is also changed. However, there is a disadvantage that the operator has to select and determine the object color vector as the characteristic value visually, and the operator's man-hour is large. In addition, it is considered that it is difficult to accurately reproduce the texture of various paints such as solid colors, metallic colors, and mica colors by visually adjusting the coefficient.

[0011] As a reference technology, [4] "Measuremen
t of Highlights in Color Images "(GJKlinker, S.
A. Shafer, T. Kanade, International Journal of Comput
er Vision, 2, 7-32 (1988)) reports a method for recognizing the color of an object itself from a color image, but the report provides a method for changing the color and texture of an image. is not.

SUMMARY OF THE INVENTION The present invention has been made in consideration of the above problems, and has as its object to provide a method of changing the color of an object image, which is capable of expressing a texture unique to the changed color when the color of the object is changed. Is to provide.

[0013]

SUMMARY OF THE INVENTION (1) The present invention relates to a method for changing the color of an object image in which the original color of the original image is changed to a changed color. A step of obtaining original color deviation characteristic data for associating the deviation parameter representing the deviation in the line of sight with the chromaticity of the reflected light in the line of sight, and using the original color deviation characteristic data to obtain the chromaticity of the original image. A bias parameter calculating step of calculating a bias parameter of each part of the object based on the reflection on the changed color plane, and a step of acquiring changed color bias characteristic data for associating the bias parameter with the chromaticity of the reflected light in the line-of-sight direction; Using the changed color bias characteristic data, determining a chromaticity corresponding to the bias parameter of each part of the object determined in the bias parameter calculating step, and determining a chromaticity of the chromaticity of the image of the changed color. Characterized by .

According to the above configuration, the bias characteristic data for associating the "bias parameter representing the bias in the viewing direction with respect to the specular reflection direction" and the "chromaticity of the reflected light in the viewing direction" is used. If the color of the reflecting surface is different, the bias characteristic data is different, that is, the state of change in the chromaticity of the reflected light according to the bias parameter is different. The difference in the manner of this change corresponds to the difference in the texture unique to each color. Accordingly, the bias parameter of each part of the original image is determined using the bias characteristic data of the original color, and then the chromaticity corresponding to the bias parameter is determined from the bias characteristic data of the changed color. This makes it possible to change the texture of the original color to the texture of the changed color at the same time as changing the original color to the changed color.

As described above, according to the present invention, by changing the color using the bias characteristic data, the texture can be changed at the same time. Therefore, the color changing process that matches the reality can be performed quickly and accurately, and it is possible to reduce the number of operation steps of the operator.

Although the present invention is suitably applied to the color change of the outer panel color of an automobile image as an object image, it is also suitably applied to the color change of another object. In particular, the effect of the present invention can be suitably obtained in a situation where the texture differs when the color of the object differs.

(2) In a preferred aspect of the color changing method according to the present invention, a lightness correcting step of correcting the lightness in the chromaticity determined in the chromaticity determining step based on the lightness of each part of the original image. In the brightness correction step, the brightness correction amount is changed in accordance with the bias parameter calculated in the bias parameter calculation step.

In the color change processing of the above (1), the gradation which is the change in lightness may not be appropriately represented in the changed color image, and this is remarkable in a portion where the bias parameter is large such as near the lower part of the object. . According to this aspect, the brightness of each part of the color-changed image is adjusted based on the brightness of the original image. Preferably, the change in the brightness of the original image is partially reflected in the changed color image by a process using a weighted average or the like. This makes it possible to appropriately represent the gradation that should originally exist. In particular, by adjusting the lightness correction amount according to the bias parameter, the gradation of the original image can be more strongly reflected in the portion where the gradation is largely insufficient, and it is possible to obtain a changed color image that appropriately expresses the gradation. Become.

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention (hereinafter, referred to as embodiments) will be described below with reference to the drawings. In the present embodiment, the exterior color of the photographed automobile image is changed using a computer graphic device. Schematically, a car image whose skin color is known is first prepared, and the declination is estimated from the brightness for each pixel. Next, the skin color of the vehicle image is changed based on the declination value and the measured spectral reflectance of another skin color.

[Declination and Deflection Spectral Reflectance] First, FIG.
The declination and the deflected spectral reflectance R used in the present embodiment will be described with reference to FIG. In FIG. 1, a light beam is emitted from a light source arranged in the incident direction L1 to a point O on the painted surface, and the viewpoint is set in the observation line-of-sight direction L2.
The color of the light reflected in the observation line-of-sight direction L2 is the color of the point O observed by the observer. The declination ξ is an angle formed between the regular reflection direction L1r and the observation line-of-sight direction L2. The deviation angle ξ is a deviation parameter representing the deviation of the viewing direction L2 with respect to the regular reflection direction L1r. In other words, declination ξ
Is a parameter representing a relative line-of-sight direction L2 based on the incident direction L1 or the regular reflection direction L1r. Declination ξ
Are positive and negative with respect to the regular reflection direction L1r, the clockwise angle (the angle approaching the incident direction L1) is positive,
The angle in the counterclockwise direction is negative.

As shown in FIG. 1, the color of the reflected light can be measured by placing the sensor in the observation line-of-sight direction L2. In the present embodiment, the spectral reflectance R (λ) is measured. The spectral reflectance R (λj) (j = 1, 2,... N) is calculated for each wavelength λ.
It represents the reflectance of j and can be used as a parameter to represent color. Hereinafter, the spectral reflectance at reflection at the declination ξ will be referred to as declination spectral reflectance R (ξ, λ).

If the reflectance is measured while moving the sensor of FIG. 1, the argument spectral reflectance R corresponding to an arbitrary argument ξ can be obtained.
(Ξ, λ) can be obtained. This measurement can be performed using, for example, a well-known goniophotometer. Further, the deflected spectral reflectance R (ξ, λ) may be obtained by calculation.

In FIG. 1, the incident direction L1 and the line-of-sight direction L2 are in the same plane. If both directions are not in the same plane, the argument ξ is defined as shown in FIG. Incident direction L
Assuming a plane formed by 1 and the line-of-sight direction L2, a straight line L3 bisecting ∠L1 · O · L2 is determined on this plane. Assuming a virtual inclined surface that gives a regular reflection arrangement relationship in the incident direction L1 and the line-of-sight direction L2, a straight line L3
Corresponds to the normal of this virtual surface. An angle ω between the straight line L3 and the vertical direction N of the painted surface is determined. Double angle of ω declination ξ
And Using the argument 偏 defined in this way, regardless of whether the incident direction L1 and the line-of-sight direction L2 are on the same plane, if the argument ξ is the same, the argument spectral reflectance R (分光,
λ) can also be considered approximately equal. Therefore, the reflectance when the incident direction L1 and the line-of-sight direction L2 are not in the same plane can be easily and accurately obtained.

[Model of appearance of outer panel color] In photographing a car, light reflected by the body is recorded by a camera. At this time, since the body has a curved surface, each part of the body has a different deflection angle. Light is reflected under conditions. In the present embodiment, the declination value of each part is estimated from the brightness of a photographed automobile image. For this estimation, a model representing the appearance of a car under realistic conditions such as outdoors is assumed.
In particular, here, a model capable of expressing that the appearance (texture) differs depending on the outer panel color is assumed as follows.

The appearance of the outer panel color of an automobile is determined by "illumination conditions", "observation conditions", "vehicle shape", and "spectral reflectance for each external angle of the outer panel color (argument spectral reflectance)". . When observing the skin color outdoors, the main factors that determine the “lighting conditions” are the irradiation direction of sunlight and ambient light, and the intensity ratio between sunlight and ambient light (relative ambient light intensity). Conceivable. The “observation condition” is the direction of the line of sight from the position of the observer's eyes to each part of the vehicle. The “vehicle shape” is defined by a normal direction of a minute surface on the body.

The relationship between the chromaticity of an arbitrary minute portion on the outer panel of the automobile and the above conditions can be considered with the model of FIG. Here, a pixel P corresponding to a certain minute portion on the outer plate
The following describes how the chromaticity of is determined by the present model.

The "declination angle with respect to sunlight" is determined from "the irradiation direction of sunlight", "the normal direction of the minute surface at pixel P", and "the line of sight to pixel P". The "illuminance of sunlight" is determined from the "irradiation direction of sunlight", the "relative intensity of sunlight", and the "normal direction of a minute surface". Then, "declination with respect to sunlight", "illuminance of sunlight", and "declination spectral reflectance"
Determines the "reflected light by sunlight." Here, FIG.
By using the reflectance measurement data described in the above section, “declination spectral reflectance R (ξ
s, λ) ”is determined, and the reflected light is specified from the reflectance and the illuminance.

On the other hand, it is assumed that the ambient light is constant regardless of the direction. The “estimated angle to the minute surface” is determined from the “normal direction of the minute surface at the pixel P” and the “viewing direction to the pixel P”. Different perspective angles result in different reflected light due to ambient light. For example, a plane perpendicular to the line of sight,
Reflected light greatly differs on a surface that is greatly inclined with respect to the line of sight. "Estimated angle of minute surface" and "Relative intensity of ambient light"
The "reflected light due to environmental light" is determined from the "declination spectral reflectance". Here, the reflected light is specified based on the assumption that the ambient light has no direction dependency. Therefore, the reflected light of the environment light is the reflected light when the ambient light of the same intensity is irradiated on the pixel P from any surrounding direction.

The "observed reflected light" is determined by synthesizing the "reflected light by sunlight" and the "reflected light by environmental light". FIGS. 4 and 5 show the relationship between declination to sunlight (hereinafter, simply referred to as declination) and brightness according to the above model. FIG. 4 shows the brightness of the solid skin color,
FIG. 5 shows the brightness of the metallic skin color. Direction of sunlight irradiation is 30 degrees (based on the direction perpendicular to the ground)
And the line-of-sight direction is set to 84 degrees. In addition, the intensity ratio between sunlight and ambient light is set to 6: 4. As described above, with respect to the positive and negative of the declination, the angle deviating in the incident direction with respect to the specular reflection direction is the positive declination.
In a typical car image, the declination is positive on the upper surface of the hood, the roof, or the like, and the declination is negative on the side surface. In other words, a transition from a positive declination to a negative declination corresponds to a transition from the top surface to the side surface.

In FIGS. 4 and 5, the thin line indicates the brightness calculated from the measured value of the deflective spectral reflectance R (ξ, λ). This measurement was performed on a flat plate of a sample using a goniophotometer according to the method described with reference to FIG. As shown, there is a large difference between the brightness based on the model and the brightness measurement. This difference
We show that this model accurately reflects the ambient light and shadow under realistic observation conditions. In this embodiment,
The chromaticity for each argument calculated by the model in FIG.
That.

[Declination Characteristic Table of Outer Plate Color] Based on the above model, an eccentricity characteristic table for associating the chromaticity of the reflected light with the declination is created. Here, an original color table and a changed color table are created. The original color is the skin color before the color is changed, and the changed color is the skin color after the color is changed. The original color table is used for estimating the argument of each part of the body from the chromaticity in the original image, which will be described later. The change color table is used to determine the chromaticity of the color change image from the argument. This table is created based on the following considerations.

(1) When photographing a car outdoors, the irradiation direction of sunlight and the relative ambient light intensity are almost constant parameters over the entire screen. On the other hand, the line-of-sight direction and the normal direction of the minute surface are parameters that change for each pixel on the screen.

Here, when changing the color of the outer panel based on the argument information, a plurality of parameters can be set for each image. However, the argument information is the only parameter in pixel units. Therefore, it is necessary to create a table capable of uniquely referring to the chromaticity value based on the irradiation direction of sunlight and the relative ambient light intensity given for each image and the argument given for each pixel. In this case, two parameters, ie, the normal direction and the line-of-sight direction of the minute surface, have to be represented by declination. In short, it is not possible to obtain two information, that is, the normal direction and the line-of-sight direction of the minute surface, from the chromaticity of the pixel. Instead, it is necessary to express both information with one parameter called the argument.

Therefore, in the present embodiment, the viewing direction is set to be constant. In order to estimate the magnitude of the effect of this setting, only the line of sight was changed in the range of 0 to 20 degrees in the horizontal direction and 78 to 90 degrees (horizontally) in the vertical direction for the metallic skin color. The color difference for each argument was determined. The range of change in the line of sight is set on the assumption that a photograph of the entire vehicle is taken. As a result, the average color difference was 0.2 and the maximum color difference was 1.5 (both ΔE * ab), and it became clear that the influence of the line-of-sight direction change was small. Therefore, in the present embodiment, a change in the line of sight is ignored, and the line of sight is represented by a constant value in both the horizontal and vertical directions.

(2) In order to estimate the declination from the brightness of each pixel of the automobile image, it is necessary that the relationship between the brightness and the declination in the declination characteristics of the skin color is one-to-one. In the estimation process, only the absolute value of the argument can be estimated, and the sign information cannot be obtained.

However, as is apparent from FIGS. 4 and 5, in the appearance of the outer panel color, the relationship between lightness and declination is as follows.
The characteristics are not always one-to-one, and the characteristics differ depending on the sign of the argument. In short, there are two declination values corresponding to one lightness, and both values are different. Therefore, the declination value cannot be determined from the lightness as it is.

In order to examine the magnitude of this effect, the color difference between the positive declination characteristic and the negative declination characteristic was investigated for four solid skin colors and 16 metallic skin colors.
The “sunlight irradiation direction”, “line of sight”, and “relative ambient light intensity” are each 30 degrees, as in FIGS. 4 and 5 described above.
84 °, 6: 4.

FIG. 6 shows the average color difference, the maximum color difference and the standard deviation of the color differences for the four solid skin colors. FIG.
Are the average color difference, the maximum color difference, and the standard deviation of the color differences in 16 metallic skin colors. From these results, there is relatively little difference between the characteristics of the metallic skin color in the positive and negative regions. Therefore, it is considered that there is no problem even if it is assumed that the positive and negative declination characteristics are the same. On the other hand, the solid outer skin color has a large positive / negative difference in a large declination region. For this reason, when performing a process on the assumption that the positive and negative declination characteristics are the same, it is considered that some countermeasure against the difference is required. This countermeasure is performed by brightness correction described later.

Here, in general, when photographing the entire vehicle, it is relatively important to reproduce the shadows on the side surfaces of the fender, the door and the like. On the other hand, on the upper surface of the hood or the like, factors other than the deflection characteristics, such as the reflection of the sky, greatly affect the appearance of the outer panel color. In consideration of this, in the present embodiment, the declination characteristics of the skin color are represented by the characteristics in the negative declination region corresponding to the appearance of the skin color on the side surface.

Based on the above (1) and (2), the conditions for creating the declination characteristic table of the outer panel color are as follows: 1) The ambient light is
2) The viewing direction is fixed at 0 ° in the horizontal direction and 84 ° in the vertical direction. 3) The characteristic in the negative declination area is defined as the declination characteristic of the skin color. , Three. Under these conditions, using the appearance model of Fig. 1,
Obtain the chromaticity value (CIELAB value) for each argument when observing the skin color in an arbitrary sunlight irradiation direction and relative ambient light intensity. Then, a declination characteristic table of the skin color that associates the declination with the chromaticity value is created. As described above, the original color table and the changed color table are created. The created table is stored in the memory of the computer graphic device. Preferably, create a table of various skin colors,
Preferably, it is stored in a memory. This makes it possible to easily and quickly carry out color changes to various outer panel colors at any time.

FIG. 8 summarizes the processing for creating the argument characteristic table. The declination spectral reflectance R (ξ, λ) is obtained by measurement or calculation (S1). “The irradiation direction of sunlight” and “the relative intensity of environmental light with respect to sunlight” are set as parameters (S2, S3).
According to the above conditions, the argument is set between 0 degrees and -180 degrees (S4). Spectral reflection light observed for each declination within this range is obtained (S5). The chromaticity (CIELAB) of each declination is obtained from the spectral reflection light, and a declination characteristic table is created (S6).

[Estimation of Declination from Car Image] First, a photograph of a car painted with the original color is prepared. Automobile photographs in catalogs and pamphlets are also available. This photograph is input to a computer graphic device using a scanner. The color values of each pixel of the input car image are converted from RGB values to CIELAB values. Then, a declination value is obtained from the chromaticity of each pixel using the declination characteristic table of the original color. Here, the argument value is obtained from the L * value of each pixel. That is, the argument value corresponding to the L * value of the pixel is read from the table. At this time, preferably, the brightness and contrast of the image are adjusted by the primary conversion. Image data (hereinafter, referred to as declination image) in which each pixel has the declination value is generated using the declination calculation value.

[External Panel Color Change Using Declination Image] Using the above declination image, a car image with a different color (changed color) from the original color is generated. Here, the declination characteristic table of the changed color is read from the memory. Then, from this table, the chromaticity corresponding to the argument value of each pixel in the argument image is read. Then, the read chromaticity is set as the chromaticity of the corresponding pixel in the image after the color change. Similar processing is performed for pixels in the entire region of the vehicle outer panel. An image having the chromaticity determined in this way is generated. This image is displayed on the display of the computer graphic device. This image is stored in the memory, and is appropriately saved in a displayable state in response to an instruction from the operator.

[Evaluation of Changed Color Image] According to the above-described color change processing, generally good results can be obtained in the case of a metallic outer panel color. However, in a portion where sunlight is not directly irradiated, such as a bumper or a lower portion of a door, gradation may be insufficient and the expression may be flat. This is considered to be the effect of creating the argument characteristic table assuming that the ambient light intensity has no direction dependency. In a situation where a car is actually observed, especially at the lower part of the car, the ambient light intensity has direction dependency, which causes gradation. On the other hand, as shown in FIG. 5, when the declination characteristic table is used, in a region where direct sunlight does not reach, that is, in a region where the declination is large, almost constant chromaticity is given, and therefore no gradation occurs. The direction dependency of the ambient light intensity changes locally depending on the three-dimensional shape and the illumination state, and it is difficult to uniquely determine the direction depending on the argument.

On the other hand, in the case of a solid outer panel color, the gradation from the front fender to the hood may be extremely insufficient. This is considered to be due to the problem described with reference to FIG. That is, in the solid skin color, the illuminance by sunlight changes according to the inclination of the minute surface, so that a consistent gradation is generated through the positive and negative regions of the declination. On the other hand, the argument characteristic table is created from the argument characteristics in the negative region. Therefore, when the argument characteristic table is used, the gradation becomes symmetric in the positive and negative regions of the argument. As a result, the brightness of the upward surface, which is originally high brightness, may be insufficient.

[Correction by Brightness of Original Image] In this embodiment, in order to solve the above-described problem, the brightness of each pixel of the original image is compared with that of the argument image and the image created from the argument characteristic table. Correction using.

First, the lightness ratio between the skin color of the original image and the skin color whose color is to be changed is determined. Here, the lightness in the chromaticity corresponding to each declination from 5 degrees to 45 degrees is obtained with reference to the declination characteristic table of each outer plate color. Then, an average value of lightness is obtained, and a ratio of the average value of the original color and the average value of the changed color is obtained.

Next, by multiplying the lightness ratio, the lightness of each pixel of the original image is converted into lightness Ly corresponding to the changed color. On the other hand, the lightness Lx of each pixel in the color-changed image created using the argument characteristic table is obtained.

The corrected brightness value Lz is determined for each pixel by a weighted average of the brightness Lx and the brightness Ly. After that, the lightness Lx of the color of each pixel of the color-changed image is replaced with the corrected lightness value Lz. Specifically, a Yx value corresponding to Lx and a Yz value corresponding to Lz are obtained, and the ratio (Yz / Y
x) is applied to the tristimulus values (Xx, Yx, Zx) of the color-changed image created using the argument characteristic table. As a result, an image whose brightness has been corrected can be obtained.

The weight coefficient in the above weighted average is a function of the argument. In a region where the argument is small, the weight of the lightness Lx (the lightness obtained from the argument characteristic) is set to be large. Conversely, in an area where the argument is large, the lightness Ly (the lightness obtained from the original image)
Is set large. Thereby, it is possible to appropriately reproduce gradation caused by uneven environmental light generated in a portion not directly irradiated with sunlight.

In the actual processing, the declination value of each pixel is obtained with reference to the declination image. And according to the above function,
A weight coefficient corresponding to each pixel is determined based on the argument value. Further, the lightness Lx and the lightness L
A weighted average of y is performed.

As a result of the above brightness correction, gradation in the lower part of the bumper and the like is preferably reproduced in a metallic color image. In addition, in a solid color image, lack of gradation is greatly improved.

[Overall Processing] FIG. 9 shows the overall processing flow. An original image of a car is input using a scanner or the like (S10). In addition, the argument characteristic table of the original color (the outer color of the car of the original image) is read from the memory of the computer graphic device (S11). With reference to this table, a declination value corresponding to the brightness is calculated from the brightness of each pixel. As the calculation result, declination image data in which each pixel has a declination value is generated (S12). S11
Similarly, the declination characteristic data of the changed color (the outer plate color after the color change) is read (S13). Then, the chromaticity corresponding to each pixel in the declination image is read from the declination characteristic table of the changed color. A changed color image in which each pixel has the chromaticity read from the table is generated (S14).

On the other hand, the lightness ratio (representative value) of the original color and the changed color is obtained (S15). Here, the declination characteristic table of both colors is referred to, and the ratio of the lightness average value is obtained.
Further, the declination value of each pixel is obtained from the declination image, and a weight coefficient for brightness correction is determined from the declination value (S16). The brightness correction of the changed color image in S14 is performed using the brightness ratio and the weight coefficient (S17). Here, a value obtained by multiplying the brightness of each pixel of the original image by the brightness ratio is obtained (Ly). This Ly
And the weighted average of the brightness Lx of the changed color image in S14 are S
This is done using 16 weighting factors. Lightness L of calculation result
z is replaced with the lightness Lx.

The preferred embodiment of the present invention has been described above. According to the present embodiment, an argument characteristic table that associates the argument with the color of the reflected light is used. The argument characteristic table shows how the color changes according to the argument. The state of this change corresponds to the texture unique to each skin color. For example, the difference between FIG. 4 and FIG. 5 corresponds to the difference in texture between the solid skin color and the metallic skin color.

Therefore, a declination is once obtained from the chromaticity of the original image using the reflection characteristic table of the original color. Next, the chromaticity of the changed color image is obtained from the argument using the changed color reflection characteristic table. Thus, at the same time as the color change, the texture generated by the original color is changed to the texture corresponding to the changed color.

According to the present embodiment, a good changed color image can be obtained even when the texture of the original color and the changed color greatly differ. For example, when changing from a solid color to a metallic color or vice versa, the texture can be changed appropriately. Therefore, it is possible to easily perform a design operation such as changing the outer skin color of a photographed automobile to another arbitrary color and examining various outer skin colors.

Further, according to the present embodiment, the color changing process (the process of accurately changing the texture together with the color) can be performed quickly and accurately with a small number of man-hours (man-hours such as visual adjustment by the operator). It is possible to do.

According to the present embodiment, the brightness of the changed color image is corrected for each pixel based on the brightness of the original image. Here, the change in the brightness of the original image is partially reflected in the changed color image. By this correction, a gradation that exists in the original image and should also exist in the changed color image is appropriately expressed. In particular, according to the present embodiment, the weight for brightness correction is changed according to the argument. This makes it possible to increase the weight of the original image and reflect the gradation of the original image more strongly in places where the gradation is likely to be large and insufficient in the changed color image before correction, which makes it possible to properly reproduce the gradation. Can be done.

A modification of this embodiment will be described. In the present embodiment, the spectral reflectance is used as the reflection parameter representing the reflected light, but the reflection parameter is not limited to the spectral reflectance. The chromaticity itself may be measured by the apparatus of FIG.

In the present embodiment, the argument is used as the argument. This declination was the difference between the specular reflection angle and the light reception angle of the illumination light. On the other hand, differently defined bias parameters or declinations with different criteria etc. may be used as appropriate. In addition, any parameter representing the light receiving direction relative to the incident direction or the specular reflection direction can be applied within the scope of the present invention.

The present embodiment is not limited to the case where the vehicle before or after the color change is painted with only one color. When the colors are classified into a plurality of outer plate colors, the car image may be divided into regions of each outer plate color, and the above-described color changing process may be performed for each region. Further, the above-described color change processing does not have to be applied to the entire vehicle body. The above processing may be applied to a part of an automobile.

In the present embodiment, the present invention is applied to a method of changing the color of an outer panel of an automobile. However, the present invention can be similarly applied to changing the color of any other object.

[Brief description of the drawings]

FIG. 1 is a diagram showing a definition of an argument.

FIG. 2 is a diagram showing a method of obtaining a declination when the incident direction and the line-of-sight direction are not in the same plane.

FIG. 3 is a diagram showing a model of how a skin color looks according to the embodiment of the present invention.

FIG. 4 is a diagram showing declination characteristics of a solid outer panel color.

FIG. 5 is a diagram showing declination characteristics of metallic skin colors.

FIG. 6 is a diagram showing a color difference between declination quadrants with respect to declination characteristics of a solid outer panel color.

FIG. 7 is a diagram showing a color difference between declination quadrants regarding a declination characteristic of a metallic outer panel color.

FIG. 8 is a flowchart illustrating a process of creating an argument characteristic table.

FIG. 9 is a flowchart showing a process of changing a skin color.

[Explanation of symbols]

L1 incident direction, L1r regular reflection direction, L2 observation line-of-sight direction, θ incident angle, ξ deflection angle.

──────────────────────────────────────────────────の Continued on front page (51) Int.Cl. ⁶ Identification code FI H04N 1/40 D 1/46 Z

Claims (2)

[Claims] An object image color changing method for changing an original color of an original image into a changed color, comprising: a bias parameter representing a deviation of a line of sight with respect to a regular reflection direction with respect to reflection on an original color plane; Obtaining original color deviation characteristic data that associates the chromaticity of the reflected light in the direction with the original color deviation characteristic data, and using the original color deviation characteristic data, a deviation parameter for obtaining a deviation parameter of each part of the object based on the chromaticity of the original image. A calculating step, and regarding the reflection on the changed color plane, a step of obtaining changed color deviation characteristic data that associates the deviation parameter with the chromaticity of the reflected light in the viewing direction, and using the changed color deviation characteristic data, A chromaticity determination step of obtaining a chromaticity corresponding to the deviation parameter of each part of the object obtained in the deviation parameter calculation step and setting the chromaticity of the image of the changed color as a chromaticity determination method. 2. The method according to claim 1, further comprising a brightness correction step of correcting the brightness in the chromaticity determined in the chromaticity determination step based on the brightness of each part of the original image. And a method of changing the color of an object image, wherein the brightness correction amount is changed according to the bias parameter calculated in the bias parameter calculation step.