JP3235151B2 - Image simulation method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for simulating a high-quality image of a color or monochrome gray image.

[0002]

2. Description of the Related Art An image is read into a computer as data.
An image simulation technology that changes the color and pattern of a specific area by using image processing technology and obtains an effect as if it is an image representing another scene is required at the design and presentation site. .

At that time, as described in JP-A-63-293676 (hereinafter referred to as known example 1),
The input color image is separated into a luminance component and a color component, arithmetic processing is performed on at least the color component, and the resultant is synthesized with the luminance component and output. Simulations that reflect shadows and the like can be performed.

Further, as described in JP-A-63-237172 (hereinafter referred to as Known Example 2) and JP-A-3-41570 (hereinafter referred to as Known Example 3), a color image inputted is disclosed. Extract the object color vector, light source color vector, etc. of the region from the color vector of the pixel of the specific region, and linearly combine the color vectors of the object color vector, the light source color vector, etc., for all the pixels in the region. By reconstructing the color vector by changing the object color vector, the light source color vector, and the like, it is possible to perform a high-quality simulation including a shadow and a reflected portion.

[0005]

The above prior art has the following problems.

(1) When a color image is separated into a luminance component and a color component, the color component is changed, and the color image is synthesized with the luminance component again,
The texture originally contained in the original image is inherited by the luminance component, and this is reflected in the final composite image.
This method is effective for applications that intend to reflect the original texture in the composite image,
It is not possible to meet the need to change the fine texture of the original image to another texture and to reflect only information such as shadows on the result from the original image. Here, information such as shadows refers to shadows that are weakened by blocking light rays corresponding to the object represented by the image, and shades caused by changes in the amount of light received depending on the angle between the light rays and the object surface. The distribution of brightness that occurs and the fact that light rays are refracted when passing through an object or scattered on an object, causing the intensity of light rays to be unevenly distributed due to the spread of sharp rays originally The resulting distribution of brightness.

(2) If the reflection of information such as shadows is limited to reference to luminance components, the following problem occurs.

(2-1) There is a possibility that information of an object color, which is independent of information such as shadows, may be included in the simulation as information such as shadows, so that it is impossible to faithfully reproduce information such as shadows.

(2-2) Since information on the color of the light source cannot be obtained, it is impossible to faithfully reproduce differences in shadows and the like caused by light sources of different colors.

(3) The application of a simulation method of separating an image into a luminance component and a color component, changing the color component, and synthesizing the color component again with the luminance component is applied to a black-and-white image. Therefore, since information other than the luminance component cannot be extracted, it is impossible to realize the information.

(4) An object color vector, a light source color vector, and the like of a specific area of the input color image are extracted from the color vector of the pixel, and the color vector is converted to the object color vector for all the pixels in the area. In the method of simulating an image by expressing by a linear combination of vectors such as light source color vector and light source color vector and changing the object color vector and light source color vector etc. and reconstructing the color vector, only the target area is To give a non-single color texture to give the object color vector of

(5) If the condition that the target region has only one object color is not satisfied, the vector cannot be extracted, so that the simulation cannot be performed on the region including the texture that is not a single color.

A first object of the present invention is to perform an image simulation for changing the color or texture of a certain area in a color or monochrome grayscale image, ignoring the influence of fine textures in the original image. An object of the present invention is to provide an image simulation method capable of reflecting a change such as a shadow.

A second object of the present invention is to provide an object simulation for changing the color or texture of a certain area in a color or black-and-white gray image while reflecting a shadow or the like in the original image. An object of the present invention is to provide an image simulation method capable of faithfully reproducing a shadow or the like without being affected by the image.

A third object of the present invention is to provide an image simulation for changing the color and texture of a certain area in a color grayscale image while reflecting the shadows and the like in the original image, and to determine the difference between the light source colors. Is to provide an image simulation method that can faithfully reproduce the image.

A fourth object of the present invention is to provide an image simulation method for changing a specific area in a color grayscale image to a texture that is not a single color while reflecting a shadow or the like or a reflection of the original image. It is in.

A fifth object of the present invention is to provide an image simulation method for changing a region including a texture which is not a single color in a color gradation image while reflecting a shadow or the like or a reflection of the original image. is there.

[0018]

According to the present invention, in order to achieve the first object, a standard color vector of an area to be simulated for an original image is determined, and further, all standard colors in the area are defined. For a pixel, the average of the color vectors is calculated in the vicinity thereof, and the ratio of this component to the standard color vector for each pixel is calculated for each pixel, and this ratio is used as information such as shading in image simulation. In addition, information such as shadows is expressed not in the luminance of a color vector but in the form of a ratio of components of a color vector and a standard color vector so that the information can be applied to a monochrome image as well as a color image. Here, the standard color vector refers to a color vector that has a distribution of color vectors similar to that of the target region, and represents a portion that is sufficiently illuminated and is considered to have no shadow or the like.

In order to achieve the second object, information such as shadows is represented not by the luminance of the color vector but by the ratio of the components of the color vector and the standard color vector.

In order to achieve the third object, information such as shadows is represented by three independent components of a color model.

In order to achieve the fourth object, when performing an image simulation for changing a specific area in a color gray-scale image while reflecting shading or the like of the original image or reflection, the object color after the change is changed. Is not made a single color in the region, but a different color vector can be obtained for each pixel.

Further, in order to achieve the fifth object, an image simulation is performed in which a region including a texture which is not a single color in a grayscale image of a color is changed while reflecting a shadow or the like or a reflection of the original image. At this time, an averaging process is performed on the original image in advance.

[0023]

According to the present invention, a standard color vector of a region to be simulated for an original image is determined, and
For all the pixels in the region, the average of the color vector is calculated in the vicinity thereof, and the ratio of each of the components to the standard color vector is calculated for each pixel. Fine textures are canceled by averaging, and the effects of shadows and the like that are independent of the textures can be extracted.Thus, it is possible to ignore the effects of the fine textures in the original image and perform image simulation that reflects the shadows and the like. Become.

Hereinafter, the operation will be described in detail.

The target image is a color image, and RGB
Think of the color model. Generally, a pixel P in an image
If the color vector of (x, y) is C (x, y), C
(X, y) is expressed as in Equation 1.

[0026]

(Equation 1)

Here, R (x, y), G (x, y) and B (x, y) are the R component of C (x, y) and G (x, y), respectively.
It is a scalar quantity representing the component and the B component.

Next, let Co be the standard color vector of the target area (Equation 2).

[0029]

(Equation 2)

Here, Ro, Go, and Bo are scalar quantities representing the R, G, and B components of Co, respectively.

Next, when averaging the color vectors in the vicinity of the pixel P (x, y) in the region to be simulated, the range of the vicinity is represented by a set of pixels ε (x, y). I do. Also, the number of pixels included in ε (x, y) is N
Expressed by (x, y).

At this time, the neighborhood ε (x, x) of the pixel P (x, y)
The average of the color vectors of y) is calculated by the vector Ca (x,
y).

[0033]

(Equation 3)

Here, Ra (x, y), Ga (x,
y) and Ba (x, y) are Ca (x, y), respectively.
Is a scalar quantity representing the R component, the G component, and the B component.
Also, Ra (x, y), Ga (x, y), and Ba (x, y)
Each component of (x, y) is represented by Equations 4, 5, and 6, respectively.

[0035]

(Equation 4)

[0036]

(Equation 5)

[0037]

(Equation 6)

Here, i is an index indicating the x coordinate, and j is y
This is an index representing coordinates.

Next, a vector F (x,
y) (Equation 7).

[0040]

(Equation 7)

Here, Rf (x, y), Gf (x,
y) and Bf (x, y) are scalar quantities representing the R, G, and B components of F (x, y), respectively. Further, Rf (x, y), Gf (x, y), and Bf (x, y)
Each component of y) is represented by Expression 8, Expression 9, and Expression 10, respectively.

[0042]

(Equation 8)

[0043]

(Equation 9)

[0044]

(Equation 10)

Here, each component of F (x, y) is represented by a pixel P
The ratio is a value of the ratio of the average color component in which the variation of the fine color component in the vicinity of (x, y) is canceled to the corresponding component of the standard color vector. Therefore, if an average color vector having a distribution of color vectors similar to that of the target region and near an area that is sufficiently illuminated and has no shadow or the like is set as a standard color vector, F (x, Each component of y) indicates the degree of shading or the like for that component at that position.

Next, an image in which the color vectors of all or some of the pixels in this area are changed is generated or obtained by inputting from outside.

The pixel P in the target area of the new image
Let the color vector of (x, y) be Cm (Equation 11).

[0048]

[Equation 11]

Here, Rm (x, y) and Gm (x,
y) and Bm (x, y) are respectively Cm (x, y)
Is a scalar quantity representing the R component, the G component, and the B component.

Here, in order to superimpose the information such as the shadow on the pixel P (x, y), a coefficient representing the information such as the shadow at that position may be applied to Cm for each component. The color vector as a result of the simulation is defined as Cr (x, y) (Equation 12).

[0051]

(Equation 12)

Here, Rr (x, y) and Gr (x,
y) and Br (x, y) are Cr (x, y)
Is a scalar quantity representing the R component, the G component, and the B component.

At this time, Rr (x, y), Gr (x,
y) and each component of Br (x, y) are given by
3, represented by Equation 14, and Equation 15.

[0054]

(Equation 13)

[0055]

[Equation 14]

[0056]

(Equation 15)

As described above, it is possible to perform an image simulation reflecting shadows and the like, ignoring the influence of fine textures in the original image.

Further, according to the present invention, information such as shadows is represented not by the luminance of the color vector but by the ratio of the components of the color vector and the standard color vector. Since the information of the object color represented by the area can be separated from the information such as the shadow, the simulation can be performed by changing the information.

Hereinafter, the operation will be described in detail.

The target image is a black and white image.

In general, the density value of a pixel P (x, y) in an image is g (x, y). Here, g (x, y) is a scalar quantity.

Next, the standard density value of the target area is set to go.
And Here, go is also a scalar quantity.

Next, when averaging the density values in the vicinity of the pixel P (x, y) in the region to be simulated, the vicinity range is represented by a set of pixels ε (x, y). I do. Also, the number of pixels included in ε (x, y) is N
Expressed by (x, y).

At this time, the neighborhood ε (x, x) of the pixel P (x, y)
The average of the density values of y) is represented by ga (x, y).

Here, ga (x, y) is a scalar quantity and is represented by Expression 16.

[0066]

(Equation 16)

Here, i is an index representing the x coordinate, and j is y
This is an index representing coordinates.

Next, the ga (x, y) and the standard density g
The value obtained by taking the ratio with o is set as F (x, y) (Equation 1)
7). Here, Fg (x, y) is a scalar quantity.

[0069]

[Equation 17]

Here, Fg (x, y) corresponds to pixel P (x, y).
The value is a ratio of an average density value in which fine density value fluctuations near y) are canceled to a standard density value.
Therefore, if the average density value in the vicinity of a portion that has the same texture as the target area and is sufficiently illuminated and has no shading or the like is set as the standard density value,
g (x, y) indicates the degree of shading or the like at that position.

Next, an image in which the density values of all or some of the pixels in this area are changed is generated or obtained by inputting from outside.

The pixel P in the target area of this new image
Let the density value of (x, y) be gm. Here, gm is a scalar amount.

Here, in order to superimpose the information such as the above-mentioned shadow on the pixel P (x, y), a coefficient representing the information such as the shadow at the position may be applied to gm (x, y). Assuming that the density value as a result of the simulation is gr (x, y), it can be written as in Expression 18. Here, gr (x, y) is a scalar quantity.

[0074]

(Equation 18)

As described above, even in a black-and-white image, it is possible to perform an image simulation in which shadows and the like are reflected while ignoring the influence of fine textures in the original image.

The simulation method for a monochrome image is equivalent to reducing the number of color components in the simulation method for a color image from three to one.

Further, since information such as shading is represented not by the luminance of the color vector but by the ratio of the components of the color vector and the standard color vector, the object represented by the target area in the color or monochrome grayscale image Since only the information such as shadows can be extracted by separating the influence of the object color, it is possible to perform an image simulation that can faithfully reproduce the shadows and the like.

Hereinafter, the operation will be described in detail.

In general, when an object is viewed only by a diffuse reflection component, as described in pages 136 to 137 of Computer Graphics Ohmsha (1987) by Eihiro Nakamae (1987), reflected light is used. Is as shown in Expression 19.

[0080]

[Equation 19]

Where Id is the intensity of the reflected light, Ip is the intensity of the parallel light, α is the incident angle, and Rd is the diffuse reflection coefficient.

Ipcos α is a portion of the light beam of Ip which hits perpendicularly to this surface, and is a value unique to the illumination and the direction between the light beam and the object surface. The fact that the intensity of the reflected light is expressed in the form of a product indicates that the degree of darkening due to a shadow or the like can be expressed separately in the reflected light in the form of a multiplication. In addition, not only in the case of darkening but also in the case of brightening, the degree can be described simultaneously with the same expression. Therefore, if the intensity of the light ray is distributed, Ip may be changed to a corresponding value. Can be represented separately. In the case of a color shading image, information such as shading is represented by Expressions 8, 9, and 10.

Each of these represents, for each component, how much the original object color appears dark due to shading or the like. Here, the object color and the degree of shading etc. are set to 2
Since the separation is performed in the form of the product of the persons, by applying the degree of the shading and the like to another object color for each component, a simulation that faithfully reproduces the shading and the like can be performed.

In the case of a black-and-white grayscale image, the same applies except that only one component is used. Further, according to the present invention, when performing an image simulation of a color grayscale image, information such as shading is represented by three independent components of a color model, so that a difference in light source color can be faithfully reproduced.

Hereinafter, the operation will be described in detail.

Since information such as shadows is represented by three independent components of a color model, as described above, information such as shadows is extracted independently of three components, information such as shadows is held, and simulation is executed. For example, it is possible to distinguish between a case where the original image is generated by hitting a red ray and a case where the original image is generated by hitting a green ray. Therefore, a simulation that faithfully reproduces the difference in light source color can be performed.

Further, according to the present invention, when simulating a specific area in a color gray-scale image while reflecting a shadow or the like, the object color after change is not made a single color in the area. Since different color vectors can be obtained for each pixel, a simulation can be performed in which the object surface represented by the region is changed to a texture that is not a single color.

Hereinafter, the operation will be described in detail.

The target image is a color image, and RGB
Think of the color model. Generally, a pixel P in an image
If the color vector of (x, y) is C (x, y), C
(X, y) is expressed as in Expression 20.

[0090]

(Equation 20)

Here, R (x, y), G (x, y) and B (x, y) are the R component of C (x, y) and G (x, y), respectively.
It is a scalar quantity representing the component and the B component.

In general, when an object is viewed only by a specular reflection component and a diffuse reflection component, there is a model in which a color vector at each point is represented by a linear combination of a light source color and an object color, as described in Known Example 2. In addition, when it is not necessarily decomposed into these two components, as described in Known Example 3,
A model in which a vector is further added to the linear combination may be considered.

According to the above model, C (x, y) is written as shown in Expression 21.

[0094]

(Equation 21)

Here, C0, C1, and C2 are object color, light source color, and variation color vectors, respectively.
k0 (x, y), k1 (x, y), and k2 (x, y)
Are coefficients applied to each color vector of the object color, the light source color, and the variation. Here, the variation vector is a cross product of the object color vector and the light source color vector.

Here, C0, C1, and C2 are written as Eqs. 22, 23, and 24, respectively.

[0097]

(Equation 22)

[0098]

(Equation 23)

[0099]

(Equation 24)

Here, R0, G0, and B0 are scalar quantities representing the R, G, and B components of the object color vector C0, respectively. Also, R1, G1, and B1
Are the R component, the G component of the light source color vector C1,
And a scalar quantity representing the B component. Also, R2, G2,
And B2 are scalar quantities representing the R component, the G component, and the B component of the color vector C2 of the fluctuation, respectively.

The above is summarized as shown in Equation 25.

[0102]

(Equation 25)

From Expression 25, Expression 26 is obtained.

[0104]

(Equation 26)

Therefore, the vectors C0, C1, and C
2, R (x, y), G (x,
y) and B (x, y), k0 (x, y), k1
(X, y) and k2 (x, y) can be calculated.

Next, changing the light source color will be considered. At this time, assuming that the light source color vector changes from C1 to C1m, the object color vector also changes to C0m as shown in Expression 27 under the influence.

[0107]

[Equation 27]

Here, R0m, G0m, and B0m are scalar quantities representing the R, G, and B components of the object color vector C0m after the light source color is changed. Also,
R1m, G1m, and B1m are scalar quantities representing the R, G, and B components of the changed light source color vector C1m, respectively.

Equation 27 can be written because the object color is affected by the light source color as incident light. Number 19
In Ip, Id is the intensity of the light beam of the corresponding component of the light source color that has not been attenuated, and Id when cosα is 1 is
This is the corresponding component of the object color.

If C0m and C1m are obtained, the cross product thereof is calculated to obtain the variation color vector C2m.
Can be requested.

As described above, the pixel P (x,
The color vector Cr (x, y) of y) is expressed as in Expression 28.

[0112]

[Equation 28]

Here, Rr (x, y), Gr (x,
y) and Br (x, y) are scalar quantities representing the R component, the G component, and the B component of the color vector Cr (x, y) of the pixel P (x, y) after the light source color change. is there. Also,
R2m, G2m, and B2m are the R component, the G component, and the B component of the variation color vector C2m after the light source color change, respectively.
A scalar quantity representing a component.

In addition, it is possible to omit the addition of the variation vector when reconstructing the vector of Expression 28, and there is no practical problem.

Next, change of the object color will be considered. Since the light source color can be handled in the same way whether it is changed or not, write C1 and
Let us represent the latest light source color.

When changing the object color, first consider the case where the latest light source color C1 is replaced with the object color already reflected, and the case where the influence of the light source color is ignored.

What is common to the above two cases is that the new object color can be treated apparently independently of the light source color thereafter. Therefore, the vector C0m (x, y) generated as a new object color or externally input may be replaced with C0 as it is. Here, the object color can take different colors depending on the position of the pixel,
This is a function of x and y.

If C0m (x, y) and C1 are obtained, the color vector C2m (x, y) of variation can be obtained by calculating the cross product thereof.

At this time, the pixel P (x,
The color vector Cr (x, y) of y) is expressed as in Expression 29.

[0120]

(Equation 29)

Here, R0m (x, y) and G0m (x, y
y) and B0m (x, y) are scalar quantities representing the R, G, and B components of the changed object color vector C0m (x, y), respectively. Also, R2m (x,
y), G2m (x, y), and B2m (x, y) are color vectors C2m (x, y) of the variation after the object color change.
This is a scalar quantity representing the R, G, and B components of y).

In addition, it is possible to omit the addition of the variation vector when reconstructing the vector of Expression 29, and there is no practical problem.

Further, as described above, even if the light source color is changed, it can be handled similarly.

Next, when the object color is changed, the light source color C
Consider a case where data is stored in the form of a diffuse reflection coefficient unique to the object surface, where 1 is not reflected.

At this time, a procedure for converting the data expressed in the form of the diffuse reflection coefficient into an object color vector reflecting the light source color C1 is required. After performing such a procedure, the object color vector C0m
If it is handled as (x, y), the subsequent procedure is the same as the above-described procedure of replacing with the object color in which the latest light source color C1 is reflected.

Here, a method of changing the data expressed in the form of the diffuse reflection coefficient to the expression of the object color vector will be described.

Since the light source color can be handled in the same manner with or without being changed, it is written as C1 to represent the latest light source color.

In general, the object color vector C0 can be written as in Equations 30 and 31 using the light source color vector C1 and the diffuse reflection coefficient vector CRd unique to the object surface.

[0129]

[Equation 30]

[0130]

(Equation 31)

Here, RRd, GRd, and BRd are each a diffuse reflection coefficient vector CRd unique to the object surface.
Is a scalar quantity representing an R component, a G component, and a B component.

Therefore, if the diffuse reflection coefficient vector is CR
When represented by d (x, y) and a function of x and y, if the object color vector of the corresponding coordinates is C0m (x, y), this is represented by Expressions 32 and 33.

[0133]

(Equation 32)

[0134]

[Equation 33]

Here, RRd (x, y), GRd (x, y)
y) and BRd (x, y) are scalar quantities representing the R, G, and B components of the diffuse reflection coefficient vector CRd (x, y) unique to the object surface, respectively.

By performing the above procedure, the color vector Cr of the pixel P (x, y) after the object color is changed is the same procedure as the case where the color is replaced with the object color in which the latest light source color C1 is reflected. (X, y) can be obtained.

As described above, it is possible to perform a simulation for changing the surface of the object represented by the area to a texture that is not a single color while reflecting a shadow or the like or a reflection.

Further, according to the present invention, when simulating a specific area in a color grayscale image while reflecting a shadow or the like or reflection, an averaging process is performed on the original image in advance. Alternatively, an area including a texture that is not a single color can also be set as a simulation target area.

Hereinafter, the operation will be described in detail.

The target image is a color image, and RGB
Think of the color model. Generally, a pixel P in an image
If the color vector of (x, y) is C (x, y), C
(X, y) is expressed as in Expression 20.

Here, C (x, y) is written as in equation 34 according to equation 21.

[0142]

(Equation 34)

Here, C0 (x, y), C1, and C2
(X, y) are color vectors of the object color, the light source color, and the variation, respectively, and k0 (x, y), k1 (x, y),
And k2 (x, y) are coefficients applied to each color vector of the object color, the light source color, and the variation, respectively. Here, the variation vector is a cross product of the object color vector and the light source color vector.

Here, C0 (x, y), C1, and C2
(X, y) are written as Equation 35, Equation 23, and Equation 36, respectively.

[0145]

(Equation 35)

[0146]

[Equation 36]

Here, R0 (x, y) and G0 (x,
y) and B0 (x, y) are scalar quantities representing the R, G, and B components of the object color vector C0 (x, y), respectively. Also, R2 (x, y), G2 (x,
y) and B2 (x, y) are scalar quantities representing the R component, the G component, and the B component of the color vector C2 (x, y) of variation, respectively.

Here, it is assumed that the object color vector is not a constant vector but can take different vectors in x and y. Therefore, the variation vector is also represented by a function of x and y. Next, when averaging processing as shown in Expressions 3, 4, 5, and 6 is performed on C0 (x, y), an appropriate ε
By setting (x, y), the object colors are averaged in the target area, and it can be treated as if it were a single color without texture.

At this time, the averaged color vector Ca
(X, y) is expressed as in Expression 37.

[0150]

(37)

Here, C0a is an averaged object color vector, C2a is an outer product of C0a and C1, and
Expressions 38 and 39 are shown.

[0152]

(38)

[0153]

[Equation 39]

The above is summarized as shown in Expression 40.

[0155]

(Equation 40)

From Equation 40, Equation 41 is obtained.

[0157]

[Equation 41]

Therefore, if the vectors C0a, C1, and C2a are obtained, Ra (x, y), G
From a (x, y) and Ba (x, y), k0 (x,
y), k1 (x, y) and k2 (x, y) can be calculated.

As described above, k0 of the pixel in the target area is obtained.
If (x, y), k1 (x, y), and k2 (x, y) are obtained, then changing the light source color or changing the object color is the same as the procedure described above. is there. At this time, it is also possible to change the object surface represented by the region to a texture that is not a single color.

As described above, it is possible to simulate a region including a texture that is not a single color while reflecting a shadow or the like and reflection.

[0161]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of an image simulation method according to the present invention will be described in detail with reference to FIGS.

In a color grayscale image as shown in FIG. 5A, it is assumed that an image as shown in FIG. 5B is generated and output by changing the color vector of a pixel in a certain target area 501. Think.

The specific processing procedure will be described with reference to the flowchart of FIG.

FIG. 2 shows a hardware configuration diagram of the present embodiment.

In step 101, a color gradation image (FIG. 5A) to be simulated is input. At this time, it is assumed that the image data has, for each pixel, information other than the color called a mask value for each pixel in addition to the color information. By giving a special mask value only to the pixels in the region where the image simulation is performed, it can be determined whether or not the pixel of interest is a target of the simulation.

In step 102, the standard color vector (Equation 2) of the image simulation target area 501 is determined and stored in the main memory in the computer 202 in FIG. Here, as the standard color vector, a color component is directly input by a numerical value from outside using the keyboard 203.

In step 103, first, the pixels of the target image are scanned, and for the pixels in the target area, the color vectors of the pixels near the target area are expressed as shown in the above-described equations (4), (5) and (6). Calculate the average (Equation 3). Here, the neighborhood ε (x, y) described in the operation is a square region with n pixels on one side centered on the target pixel 301. The value n, which is the size 302 of the average area, is directly input numerically from outside using the keyboard 203. The value of the square of n is substituted for N (x, y) described in the operation. Next, as shown by the above-described equations 8, 9, and 10, the ratio (equation 7) of each component to the standard color vector is calculated for each pixel, and this is calculated in the main memory in the computer 202. To hold.

In step 104, the computer 202
, Another image in which the color vector of the target area 501 is changed is generated. This is the keyboard 203
Is generated by directly inputting the color components by numerical values using, and filling the area with the color vector. Here, the first image is referred to as a first image, and the image generated here is referred to as a second image.

In step 105, the pixels of the first image are scanned, and the pixels in the target area are subjected to the second operation at the same position as the pixels, as shown by the equations (13), (14) and (15). A new color vector (Equation 12) is generated by multiplying each component of the color vector of the pixel of the image by the value of the above ratio corresponding to this position, and the color vector of the pixel of the first image is calculated by To generate a third image. At this time, the pixels outside the target area
The color vector of the pixel of the image is adopted as it is. next,
This image is displayed on the CRT 201 and the external storage device 205
Is output as image data.

According to the above-described procedure, even when the target area of the first image is not a single color but contains fine textures and the like, ignoring those effects, only the information such as shadows is ignored. Is extracted, and the area is changed to another color and then applied, thereby enabling an image simulation which is not affected by the previous texture and reflects a shadow or the like.

The embodiment described above can be modified as follows.

First, the standard color vector determined in step 102 and the ratio calculated in step 103 are not stored in the main memory in the computer 202, but are stored in the external storage device 205. The ratio used in the external storage device 205
It is also possible to carry out the modification in such a way that it is obtained by calling out a certain ratio. By changing this embodiment, various types of data can be held in the external storage device, and a combination of the first image with information such as shadows different from the generated information such as shadows is simulated. It is possible to do.

Second, the standard color vector in the step 102 is determined not by directly inputting the color components numerically from the outside using the keyboard 203 but by any one of the following five methods. Changes are possible. The first method is to display a color grayscale image on the CRT 201, designate a pixel in a bright and non-shaded area in the target area of the image using the mouse 204, and use the color vector as a standard color vector. It is a method. The second method is to display a color grayscale image on the CRT 201, designate a pixel in a bright and non-shaded area in the target area of the image using the mouse 204, and display n pixels on one side in the vicinity thereof. This is a method in which a square area is taken, and the average of the color vectors of the pixels in the square area is obtained by the same procedure as in the above-described step 103, and is used as a standard color vector. At this time, the magnitude of n is given from outside.
A third method is to display a color grayscale image on the CRT 201, designate a pixel in a bright and non-shaded area in the target area of the image using the mouse 204, and display n pixels on one side in the vicinity thereof. This is a method of taking a square area, averaging the color vectors of the pixels in the square area and the pixels in the target area, and using the average as a standard color vector. At this time, the magnitude of n is given from outside.
Also, N (x, y) required for averaging is 2 of n.
It is not the power but a value obtained by counting the target pixels. The fourth method is to display a color grayscale image on the CRT 201, designate a pixel in a bright, non-shaded area in the target area of the image using the mouse 204, and display n pixels on one side in the vicinity thereof. Take a square area and calculate the average of the color vectors of the pixels in it. Here, n is made variable and n
Is sequentially increased by two from 1 and when the average value converges for all components, a color vector having the converged value as a component is used as a standard color vector.
Here, in the convergence determination, a determination value is appropriately determined, and if the fluctuation is smaller than that, it is assumed that the convergence has occurred. The fifth method is to display a color grayscale image on the CRT 201, specify a pixel in a bright, non-shaded area in the target area of the image using the mouse 204, and display n pixels on one side in the vicinity thereof. Take a square area, calculate the average of the color vectors for the pixels in the square area and the pixels in the target area, where n is variable,
In this method, n is increased by 2 in order from 1 and, when the average value converges for all components, a color vector having the converged value as a component is used as a standard color vector. Here, in the convergence determination, a determination value is appropriately determined, and if the fluctuation is smaller than that, it is assumed that the convergence has occurred. Also, N (x, y) required for taking the average is n
Instead of the square of the target pixel.

Third, in step 103, the method of calculating the average of the color vector can be changed to any of the following five methods. The first method is to average the color vectors of the pixels in the square area and the pixels in the target area.
At this time, N (x, y) required for taking the average is n
Instead of the square of the target pixel.
The second method is a method in which, when a square area is used in obtaining the standard vector in step 102, the size of one side of the square area is used as it is as the size n of the average area. Of course, this is a method that can only be implemented when step 102 has been modified and implemented with a method using square regions. Similarly, in the case where a square region is used in obtaining the standard vector in step 102, the size of one side of the square region is used as it is as the average region size n. Here, in addition to this, pixels within the square area, and
The average of the color vectors is determined for the pixels included in the target area. Of course, this is a method that can only be implemented when step 102 has been modified and implemented with a method using square regions. Also, at this time, N (x, y) required for averaging is not the square of n but a value obtained by counting the target pixels. The fourth method is to take a square area of n pixels on the side in the vicinity of the target pixel and calculate the average of the color vectors of the pixels in the square area. In this method, when the average value converges for all components, a color vector having the converged value as a component is used as an average color vector. Here, in the convergence determination, a determination value is appropriately determined, and if the fluctuation is smaller than that, it is assumed that the convergence has occurred. The fifth method is to take a square area with n pixels on the side in the vicinity of the pixel of interest and calculate the average of the color vectors for the pixels in the square area and the pixels in the target area. Here, n is variable,
This is a method in which n is increased by two in order from 1 and when the average value converges on all components, a color vector having the converged value as a component is used as an average color vector. Here, in the convergence determination, a determination value is appropriately determined, and if the fluctuation is smaller than that, it is assumed that the convergence has occurred. Also, N (x, y) required for taking the average is n
Instead of the square of the target pixel.

Fourth, in step 104, the method of generating the second image can be changed to any of the following three methods. The first method is to allow an arbitrary color vector to be obtained instead of filling the target area with the same color. Here, color vectors are generated interactively using the mouse 204 and the keyboard 203. In the second method, a new fourth image (FIG. 4A) is read, and a second image is generated by texture-mapping the new fourth image to the target area 401 of the first image (FIG. 4).
(B)) method. Thereby, an image simulation when a favorite image is pasted on the target region of the first image can be performed. A third method is that, in the second method, after reading the fourth image, a fifth image in which global shades are removed from the fourth image is generated, and the fifth image is set as an object of the first image. This is a method of generating a second image by performing texture mapping on an area. This allows
Unnecessary shading that was originally in the fourth image can be prevented from affecting the final third image. It should be noted that the following procedure is taken to remove the shading. First, the method of determining the standard color vector in step 102 or the above-described modification method is performed on the fourth image. Next, a method of obtaining the ratio in step 103 or the above-described modification method is performed on the fourth image. Next, for each pixel of the fourth image, the value of each component of the color vector is divided by the corresponding ratio value. As a result, an image in which shading and the like have been removed from the fourth image is generated. This may be used as the fifth image.

Fifth, in generating the second image in step 104, the pixels changed from the first image are
Not all the pixels in the target area of the first image may be part of them. That is, it is possible to change the color vector for a subset of the target area of the first image so as to change the color vector. At that time, a method of filling with a single color as shown in step 104 is also possible, and all the above-mentioned modifications are also feasible.

Sixth, a method of generating the second image in step 104 will be described. This step is taken offline as a procedure, and the result is stored in the external storage device 2.
05 and store it in the external storage device 2 when using it.
05 can also be read. At this time, the method of generating the second image may be the same as the method of step 104 described above, or may be replaced with any of the above-described modifications. By changing this embodiment, various types of data can be held in the external storage device, and an image simulation can be performed on the first image in combination with various second images. .

Seventh, in step 105, a third image is generated, this image is displayed on the CRT 201, and output as image data to the external storage device 205. After the generation of the third image, The processing can be ended only by displaying the image on the CRT, or can be ended by outputting the image data without displaying the image on the CRT. Further, the output of the image data can be changed not to the external storage device but to a hard copy device or another arbitrary device via a network.

Eighth, in the above embodiment, the color model is converted to RGB.
Although the description has been given as a color model, the present invention can be similarly implemented with an arbitrary color model.

Ninth, the above-described embodiment is directed to a color grayscale image, but as described in the operation, it can be applied to a black-and-white grayscale image.

Next, another embodiment will be described in detail with reference to FIG. 6 and FIGS. 2 to 5 which have already been used.

The specific processing procedure will be described with reference to the flowchart of FIG.

The hardware configuration diagram shown in FIG. 2 can be used for explaining the present embodiment.

In step 601, a color gradation image (FIG. 5A) to be simulated is input. At this time, it is assumed that the image data has, for each pixel, information other than the color called a mask value for each pixel in addition to the color information. By giving a special mask value only to the pixels in the region where the image simulation is performed, it can be determined whether or not the pixel of interest is a target of the simulation.

In step 602, an object color vector and a light source color vector are extracted from the target region 501 of the image simulation by the same method as in the known example 2 or the known example 3. In addition, as described in the operation, a color vector of variation is generated from these. Next, these vectors are stored in the main memory in the computer 202 of FIG.

In step 603, first, the pixels of the target image are scanned, and for the pixels in the target area, the coefficient for expressing the color vector of each pixel by the linear sum of the above vectors is determined by the procedure described in the operation. I do. Next, this is stored in the main memory in the computer 202.

At step 604, the computer 202
, Another image in which the color vector of the target area 501 is changed is generated. This is an external keyboard 20
3. Generated by interactively specifying a pixel in the image and its color vector using the mouse 204. Here, the first image is referred to as a first image, and the image generated here is referred to as a second image.

In step 605, the pixels of the first image are scanned, and for the pixels in the target area, as indicated by the operation, the color vector of the pixel of the second image at the same position as the pixel is set to the new object color. To generate a color vector corresponding to this position, and replace the color vector of the pixels of the first image with this to generate a third image. At this time, the color vector of the pixel of the first image is directly used for the pixel outside the target area. Next, this image is displayed on the CRT 201 and output to the external storage device 205 as image data.

According to the above-described procedure, a simulation can be performed in which the surface of the object represented by the area is changed to an arbitrary texture that is not a single color while reflecting shadows and the like and reflection. The embodiment described above can be modified as follows.

First, before extracting the object color vector and the light source color vector in step 602, it is possible to add a process of averaging the target image.
As a result, an image simulation can be performed even in a region including a texture instead of a single color in the image as shown in the operation. Here, for the image averaging process, any of the following methods is possible. The first method is a process of taking a square area of n pixels on each side as shown in FIG. 3 for each pixel in the target area, and averaging the color vectors of the pixels included in this area. At this time, the magnitude of n is given from outside. The second method is a process of similarly taking a square area of n pixels on one side for the pixels in the target area, and averaging the color vectors of the pixels included in this area and the pixels in the target area. At this time, the magnitude of n is given from outside. In addition, the number of pixels involved in the averaging, which is necessary when taking the average, is not the square of n but a value obtained by counting the number of pixels involved.

Second, in step 604, the method of generating the second image can be changed to any of the following three methods. The first method is limited to the case where the target area is not a single color but includes a texture, but this area is filled with the same color using a color vector specified externally using the keyboard 203. In the second method, a new fourth image (FIG. 4A) is read, and the second image is generated by texture-mapping the new image to the target area 401 of the first image (FIG. b)) method. Thereby, an image simulation when a favorite image is pasted on the target region of the first image can be performed.
A third method is that, in the second method, after reading the fourth image, a fifth image in which global shades are removed from the fourth image is generated, and the fifth image is set as an object of the first image. This is a method of generating a second image by performing texture mapping on an area. Thus, unnecessary shading that was originally in the fourth image can be prevented from affecting the third image as the final result. It should be noted that the following procedure is taken to remove the shading. First, the method of determining the standard color vector in step 102 or the above-described modification method is performed on the fourth image. Next, a method of obtaining the ratio in step 103 or the above-described modification method is performed on the fourth image. Next, for each pixel of the fourth image, the value of each component of the color vector is divided by the corresponding ratio value. As a result, an image in which shading and the like have been removed from the fourth image is generated. This may be used as the fifth image.

Third, steps 603 and 604
During this time, a procedure for changing the light source color can be added. Here, the keyboard 203 and the mouse 2 are externally
04 is used to interactively designate a light source color vector. As a result, as described in the operation, an image simulation in which the light source color is changed can be performed.

Fourth, at the stage of generating the second image in step 604, as described in the operation, a change can be made to reflect the latest light source color. Here, if the light source color has not been changed, the light source color is the light source color extracted in step 602. As described above, a procedure for changing the light source color is added between step 603 and step 604, and the light source color is changed. Is the light source color after the change. As described in the operation, each component of the color vector reflecting the light source color is obtained by a procedure of multiplying each component of the first given color vector by each component of the light source color. As described above, since the object color can be set as affected by the light source color, more realistic image simulation can be performed.

Fifth, in generating the second image in step 604, the pixels changed from the first image are
Not all the pixels in the target area of the first image may be part of them. That is, it is possible to change the color vector for a subset of the target area of the first image so as to change the color vector. At that time, the method shown in Step 604 is also possible, and all of the above-mentioned modifications are also feasible.

Sixthly, at the stage of extracting the light source color vector and the object color vector in step 602 and holding them, instead of holding them in the main memory of the computer, they are held in the external storage device 205. Further, at a stage where these vectors are required in the subsequent processing, it is also possible to carry out such a modification that the vectors are obtained by calling them from an external storage device. By changing this embodiment, various types of data can be held in the external storage device, and an arbitrary vector set can be extracted from various data to perform image simulation.

Seventh, a method of generating the second image in step 604 will be described. This step is taken offline as a procedure, and the result is stored in the external storage device 2.
05, and store it in the external storage device 2 when using it.
05 can also be read. At this time, the method of generating the second image may be the same as the method of step 604 described above, but may be replaced with any modified example described above. By changing this embodiment, various types of data can be held in the external storage device, so that an image simulation can be performed on the first image in combination with various second images.

Eighth, in step 605, a third image is generated, this image is displayed on the CRT 201, and output to the external storage device 205 as image data. After the generation of the third image, The processing can be ended only by displaying the image on the CRT, or can be ended by outputting the image data without displaying the image on the CRT. Further, the output of the image data can be changed not to the external storage device but to a hard copy device or another arbitrary device via a network.

Ninth, the above embodiment uses color models of RGB.
Although the description has been given as a color model, the present invention can be similarly implemented with an arbitrary color model.

[0199]

According to the present invention, the following effects can be obtained.

(1) When performing an image simulation for changing the color or texture of a certain area in a color or black-and-white gray image, the influence of fine texture in the original image is ignored, and changes in shadows and the like are reflected. A possible image simulation can be performed.

(2) When performing an image simulation for changing the color or texture of a certain area in a color or black-and-white shaded image while reflecting a shadow or the like in the original image,
It is possible to perform an image simulation that can faithfully reproduce shadows and the like without being affected by the object color.

(3) When performing an image simulation in which the color or texture of a certain area in a color shading image is changed while reflecting a shadow or the like in the original image, an image capable of faithfully reproducing a difference in light source color. A simulation method becomes possible.

(4) It is possible to perform an image simulation in which a specific area in a color shading image is changed to a texture that is not a single color while reflecting a shadow or the like or a reflection of the original image.

(5) It is possible to perform an image simulation in which a specific area including a texture that is not a single color in a grayscale image of a color is changed while reflecting a shadow or the like or a reflection of the original image.

[Brief description of the drawings]

FIG. 1 is a flowchart showing a processing procedure of an embodiment of an image simulation method according to the present invention.

FIG. 2 is a hardware configuration diagram.

FIG. 3 is a diagram for explaining a method of averaging color vectors around a target pixel;

FIG. 4 is a diagram for explaining a method of generating an image by texture mapping.

FIG. 5 is a diagram for explaining the purpose of the present embodiment.

FIG. 6 is a flowchart showing a processing procedure of another embodiment of the image simulation method of the present invention.

[Explanation of symbols]

201 CRT, 202 computer, 203 keyboard, 204 mouse, 205 external storage device, 30
1. Target pixel, 302: size of average area, 401: target area, 501: target area.

──────────────────────────────────────────────────の Continued on the front page (72) Inventor Makoto Kato 1099 Ozenji Temple, Aso-ku, Kawasaki City, Kanagawa Prefecture Inside Hitachi, Ltd.System Development Laboratory Co., Ltd. (72) Shinichiro Miyaoka 1099 Ozenji Temple, Aso-ku, Kawasaki City, Kanagawa Prefecture (56) References JP-A-3-41570 (JP, A) JP-A-63-237172 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06T 11/60 120 G06T 1/00 510 G06T 7/00 100 JICST file (JOIS)

Claims (2)

(57) [Claims] 1. A method of performing an image simulation of a region including a texture to the color or black-and-white grayscale image, enter the first image, including texture to be simulated in the first image defining a standard color vector region, for all the images of the area including the texture, calculates the average value of the color vector in the vicinity, the mean value and the previous
Calculates the ratio of each color component to the standard color vector for each pixel
And the color of at least a portion of the pixels in the region including the texture
Enter the second image vector is changed for every pixel of the area including the texture, the color vector of the pixel of the second image corresponding to the image, each component the ratio corresponding to that pixel Color vector
An image simulation method comprising: generating a third image including a torque, and outputting the generated third image . 2. A method of performing an image simulation of a region including a texture to the color or black-and-white grayscale image, enter the first image, of the object region to be simulated in the first image defining a standard color vector, for all the images of the target area, calculates the average value of the color vector in the vicinity thereof, the ratio of each color component of the standard color vector and the mean value calculated for each pixel, the The color vector of at least some pixels in the target area is changed.
Enter the further the second image was, for all the pixels of the target area, the color vector of the pixel of the second image corresponding to the image, was allowed to act the ratio corresponding to the pixel for each component An image simulation method comprising: generating a third image including a color vector; and outputting the generated third image .