JPH07203299A - Color adjustment for color picture and method and device for composing picture - Google Patents

Abstract

PURPOSE:To obtain a smooth change in color similar to the case with an original picture by applying color mixture processing to picture element data at a position in the case where a border area is wide accordingly while using an object color as a background color. CONSTITUTION:An object desired to be extracted by-an object extract section 3 is commanded to a picture displayed on a color display device 8 or the like and an object extract section 6 obtains a picture of the object. Then a picture synthesis section 10 detects a direction of the object and background with respect to each surrounding element of the extracted object and decides the object color and the background color based on the detected direction succeedingly. Furthermore,the ratio of color mixture of the surrounding picture element is calculated based on the object color and the background color to be decided and on the picture element of each position. Data of surrounding picture elements are calculated by applying color mixing processing to a new object color and the original background color by using the color mixture ratio. Thus. a natural border similar to that of the original picture is obtained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention extracts an arbitrary object in an image, changes the color, shape, size, etc. of the extracted image and superimposes it on the original image or another background image ( Synthetic)
The present invention relates to a color image color adjusting and image synthesizing method for creating a new image, and a color adjusting and image synthesizing apparatus.

Change the color of an object (color change), change the size and shape of an object (enlarge / reduce / deform), synthesize an object with another image, and recognize an object (person, car, desk, etc.) Various processing of color images, such as computer graphics (C), product design systems and computer graphics (C
It is used in a wide range of fields such as creating G), creating a document with pictures (desktop publishing), creating an original for printing.

For example, in the design of products (packages, cars, etc.) and posters, a computer creates a CG image, or a reading device inputs a natural image, especially a color image, to obtain a specific object. The image is extracted, and the size and shape of the extracted image are changed, and the color and background are changed. In the use of color images on such a computer, there is a high-precision and simple method for extracting objects, changing colors, enlarging / reducing / deforming, synthesizing, recognizing processing methods, and an easy-to-use computer-human interface (human Interface) is required.

In particular, the above-mentioned color change, enlargement / reduction / deformation,
When an object image that has undergone various processing such as composition and recognition is overlaid (combined) with the original image or another background image, the natural color change at the boundary between the object and the background is saved, and the original image is saved. It is important to create a natural composite image similar to.

[0005]

12 and 13 are diagrams for explaining a conventional method for detecting an object color and a background color. In the conventional superimposition or combination of various processed images such as color change, enlargement / reduction / deformation, composition, recognition, etc. with the original image or other background images, is there any processing of the object and background boundary part? Alternatively, the operator performs processing such as blurring the boundary portion while viewing the image on a display or the like.

[0006]

In the conventional method which does not process the boundary portion of the object and the background, the quality of the image (naturalness of the boundary) is worse than that of the original image (the boundary is unnatural,
It becomes dirty and becomes emphasized). Further, there is a problem that no matter how much time is spent, it will not be successful unless the user has processing experience and knowledge about colors.

In order to improve this, the applicants of the present invention have disclosed a boundary processing method for preserving the natural color change of an object / background boundary when superimposing (combining) a processed image on a background image. It is disclosed in Japanese Patent Publication No. 05-120417 "Color image color adjustment and image composition method and apparatus". Figure 1 above
2, the method for detecting the object color and the background color is the method disclosed in the above-mentioned Japanese Patent Laid-Open No. 05-120417, "Color image color adjustment and image composition method and apparatus".

First, pixels around the extracted object (boundary pixels)
With respect to positions {eg, in FIG.
As shown in FIG. 13, two pixels apart, vertically and horizontally
(See Figure 12 (a), (b)), Diagonal direction (See Figure 13 (c))
Pixel position data in each direction} of the object color (Ro, Go, Bo)
And the background color (Rb, Gb, Bb) is detected. Next, using these, the ratio K of the color mixture at the original image boundary is obtained based on the additive color mixture formula shown in FIG. Specifically, the pixel value at each position of the boundary, the object color (Ro, Go, Bo) and the background color (Rb, G
b, Bb), the ratio K of color mixture at each position is obtained by the above additive color mixture equation. Furthermore, the object color and the background color after the color change, or the original object color and the background color of another background image,
A new boundary is created by mixing the color mixture ratio K obtained earlier.
(Around object) Obtain pixel data (R, G, B).

The above boundary processing method is effective for most images and can realize a natural boundary similar to that of the original image. However, this method has a problem that the processing accuracy becomes poor for objects with unclear backgrounds and backgrounds, and for areas with a wide boundary color change area (when there is a wide shadow). It was That is, if there is little change in color,
As shown in Fig. 12 (a), (b), and Fig. 13 (c), if the pixel at the fixed position is the object color and the background color, the ratio K of the color mixture becomes an inaccurate value due to the influence of noise. There was something that happened.

In view of the above-mentioned conventional drawbacks, the present invention extracts an arbitrary object in an image and further changes the color, shape, size, etc. of the extracted image to obtain the original image and other background images. Create a new image by overlaying (compositing) on a color image. In the image composition method, when the boundary between the object and the background is unclear, or when the boundary color change area is wide. Another object of the present invention is to provide a color image color adjusting method, an image synthesizing method, and an apparatus capable of performing color mixing processing with high accuracy.

[0011]

1 and 2 are diagrams showing the principle of the present invention. The above problems can be solved by a color adjusting method for a color image, an image synthesizing method, and an apparatus configured as follows.

(1) Extract (100) any object in the image,
The color of the extracted object is changed (101), the peripheral portion of the color changed object is subjected to color mixing processing with the background image boundary in the image, and the color adjustment method of the color image is superimposed (102), As a method 102 for superimposing the color-changed object on the original background image, with respect to the peripheral pixels of the extracted object, the directions of the object side and the background side are detected (1020), and the object color corresponding to the peripheral pixels is detected. , The original background color and the object color after the color change are determined (1021), and the data value of a new surrounding pixel is obtained using the object color, the background color, and the object color after the color change
Configure as (1022,1023).

(2) Extract (100) an arbitrary object from the image,
Predetermined image processing is applied to the extracted object to obtain a color,
In the image synthesizing method of color images, the shape and size are changed (101a), the peripheral portion of the image processing object is subjected to color mixing processing with an arbitrary background image boundary, and the image processing is performed in the color image combining method. As a method 102a for superimposing the modified object subjected to the above on an arbitrary background image, the direction of the object side and the background side is detected with respect to the surrounding pixels of the extracted object (1020), and the object color corresponding to the surrounding pixels And an arbitrary background color and the background color of the background image are determined (1024), and the data value of a new surrounding pixel is obtained using the object color, the background color, and the background color of the background image (1022). , 1023).

(3) As a method 1020 for detecting the directions of the object side and the background side with respect to the surrounding pixels of the extracted object described in the above items (1) and (2), a constant distance from the surrounding pixels is used. A plurality of sets of pixels separated from each other is obtained, the difference between the data of the obtained plurality of sets of pixels is calculated, and the direction of the set having the maximum difference is the direction of the object and the background. Alternatively, a normal line to the extracted object shape at the positions of the surrounding pixels is detected, and the detected normal direction is recognized as the direction of the object and the background.

(4) In the above (1), (2),
A method of determining the object color and background color corresponding to the surrounding pixels, and the object color after the color change or the background color of the background image.
As the above, the difference between the data of adjacent pixels of a plurality of pixels in the direction of the detected object and the background is calculated, and the calculated difference value is detected in advance at a pixel at a position smaller than a preset value. The object color and the background color, and the object color after the color change, or the background color of the background image is determined based on the selected pixels. Alternatively, a straight line passing through a plurality of pixels in the direction of the detected object and background is detected, and the difference between the value determined by the straight line and the data value of the pixel in the object and background direction becomes larger than a preset value. The pixel at the position is detected, and the object color and the background color, and the object color after the color change, or the background color of the background image is determined based on the detected pixel. Alternatively, from a pixel value of a pixel that passes through a plurality of pixels in the direction of the detected object and background, a curve that approximates a change in the pixel value is detected, and an object color is determined based on the pixel value of the pixel determined by the curve. The background color, the object color after the color change, or the background color of the background image is determined.

(5) A device for performing color adjustment of a color image and image composition, which specifies a memory means 2 for storing input image data such as RGB signals and an arbitrary object in the input image, Object extracting means 3 and 5 for extracting an object from the designated result, image processing means 9 for changing the color, shape, size, etc. of the extracted object, and the image processing result as an original image or other It is composed of an image synthesizing means 10 which superimposes on a background image and, at the same time, performs color mixing processing of an object and a background boundary.

[0017]

[Function] The cause of the above problem is as shown in FIG. 13 (d).
This is because the object color and background color detection accuracy of the additive color mixture method shown in (3) is poor. Specifically, the pixels around the extracted object
Positions that are a fixed distance apart from (boundary pixel),
For example, in the above-mentioned example, the vertical,
Object color (Ro, Bo, Go) with pixel values of horizontal and diagonal pixels
And the background color (Rb, Gb, Bb).

Therefore, in the present invention, regarding the boundary of the object in the image, when the boundary area is wide (the color change is wide), the data of the pixel at the position corresponding to the boundary color is set as the object color and the background color, By using these to perform color mixing processing on the boundary between the object after the color change and the original image, or at the boundary between the object after the image processing and another background image, a smooth color change similar to the original image is realized. The object / background boundary is obtained.

As described above, when the boundary area is wide (the change in color extends over a wide range), the data of the pixel at the corresponding position is set to the object color and the background color, so that the unclear boundary area is Accordingly, the pixel value of the pixel at the position without ambiguity can be detected as the object color and the background color, so that it is possible to realize highly accurate color mixing processing for any boundary. is there.

[0020]

Embodiments of the present invention will be described in detail below with reference to the drawings. FIG. 1 and FIG. 2 described above are principle configuration diagrams of the present invention.
3 to 11 are views showing an embodiment of the present invention, and FIGS. 3 and 4 show a case where the direction of an object and the background is obtained by detecting the difference between pixel values in a predetermined direction. FIG. 5 shows the case where the direction of the object and the background is the direction of the normal line of the straight line passing through the contour pixels, and FIGS. 6 and 7 show the case where the boundary pixel ( 8 shows a case in which the difference between the data of adjacent pixels is sequentially calculated from the pixels on the contour) in the above determined direction, and the value of the pixel at the position where the difference becomes smaller than a predetermined threshold value is used. FIG. 9 shows that the tangent line of the boundary pixel in the determined direction is calculated, and the difference between the calculated tangent line and the actual pixel value is calculated. FIG. 10 shows a case where a pixel is used, and FIG. 10 shows a change in the pixel value of the pixel in the determined direction. Then, a curve (for example, a cubic curve) that approximates the change in the pixel value is obtained, and for example, the case where the pixel at the position where the pixel value converges is used from the approximate curve is shown. FIG. The example of a structure of a color adjustment and an image synthesizer is shown.

According to the present invention, an object to be extracted is designated in the displayed image to obtain an object image, the color of the extracted object is changed, and the original image or another background image is superimposed. When matching, for each surrounding pixel of the extracted object, the object and background are not along the contour of the extracted object, for example,
Vertical, horizontal, diagonal, or a direction such as a tangent normal line of a boundary pixel is detected, and in the detected direction, for example, a difference between data of adjacent pixels is sequentially calculated, and the difference value is calculated in advance. The pixel value of the pixel at a predetermined position is set to the object color or the background color according to the state of the boundary area, for example, the pixel at the position smaller than the set reference value is set as the object color or the background color. Is the method necessary for carrying out the present invention. The same reference numerals indicate the same objects throughout the drawings. Hereinafter, FIG. 1, FIG. 2, and FIG.
The color adjusting method of the color image, the image synthesizing method, and the apparatus thereof according to the present invention will be described.

FIG. 1 is a diagram for explaining the flow of the principle of color adjustment processing of a color image, which corresponds to claim 1 of the present invention. An object image is obtained by pointing an object to be extracted to an image displayed on a display or the like. Next, the color of the extracted object is changed and superimposed on the original image. At the time of superimposition, processing of the object and background boundary portion is performed. {Refer to the processing steps 100, 101, 102 of FIG. 1} First, with respect to each peripheral pixel of the extracted object, a certain direction between the object and the background (a direction not along the outer shape of the extracted object) is detected. Specifically, the direction is any of up, down, left, right, and diagonal. Subsequently, the object color and the background color are determined based on the detected direction.
Furthermore, the color mixture ratio K of surrounding pixels (object / background boundary) is
Based on the pixel value at each position and the determined object color and background color, calculation is performed by the above-described additive color mixture formula (see FIG. 13 (d)).

Using this color mixture ratio K, the new object color is mixed with the original background color to calculate the data of the surrounding pixels, and a natural boundary similar to the original image is realized. {Fig. 1
Processing steps 1020 to 1023 of FIG. 2} FIG. 2 is a diagram for explaining the principle flow of the color image synthesizing process, which corresponds to claim 2 of the present invention. An object image is obtained by pointing an object to be extracted to an image displayed on a display or the like. Various kinds of image processing such as changing the color or shape of the extracted object or changing the size of the extracted object are performed, and the synthesis processing is performed on another background image. At that time, similar to the color adjustment processing of FIG. 1, the color mixing processing of the object / background boundary portion is performed. In addition, in a series of processing,
In some cases, various processes such as color and shape changes are not performed, and the object is simply combined with another background image.

FIGS. 3 and 4 are one of the methods for detecting the directions of the object side and the background side with respect to the surrounding pixels (hereinafter referred to as boundary pixels) of the extracted object, which correspond to claim 3 of the present invention. It is a figure explaining.

In order to perform the color mixing process, first, the boundary pixel to be targeted in FIG. 3 (specifically, the pixel on the contour line, for example, the Japanese Patent Application No. 04-275593, which the applicant of the present invention has previously filed. With respect to the pixels on the contour line obtained by extracting the contour line by the method disclosed in the publication "Object contour detection method, object extraction method and image processing apparatus", which direction (up and down in the figure) , Left, right, diagonal) inside the object,
Also, it is judged whether there is an outside. Based on this, the object color and background color used in the calculation of color mixing processing are obtained. (How to obtain the object color and the background color will be described later) As a method of detecting the directions of the object side and the background side, a constant distance is used in the present embodiment, and from the target boundary pixel in the examples of FIGS. 3 and 4. For example, the difference of the data of the pixels at the positions separated by two pixels is calculated. This calculation is performed, for example, in the vertical direction (Y direction), the horizontal direction (X direction).
Direction) and two oblique directions (Z1 and Z2 directions), for a total of 4 directions. The following method can be considered as this calculation method.

1) Each component of color data (RGB or YC
For each rCb etc.) 2) Calculate the brightness component (R + G + B or Y). In the method 1), three results (RGB, etc.) are obtained for one direction. Therefore, in 4 directions, 12 results are obtained. For example, the direction including the maximum value of these 12 results is set as the detection result of the direction of the object side and the direction of the background side. In the method 2), the maximum direction of the four is detected.

Comparing the methods 1) and 2), the method 1) has a small amount of calculation and is simple. The method 1) can handle cases where the brightness is the same but the colors are different, and is more accurate than the method 2).

Regarding the directions of the object side and the background side, the two directions of Z may be omitted and the two directions of XY may be omitted, but the accuracy may be somewhat deteriorated, but the amount of calculation can be reduced. FIG. 5 is a diagram corresponding to the invention described in claim 4 of the present invention, and is a diagram for explaining a second method for detecting the directions of the object side and the background side with respect to the surrounding pixels of the extracted object. The present embodiment is a method of obtaining the directions of the object and the background by obtaining the normal line to the continuous boundary pixels (outline of the object) around the object. A concrete method is shown in FIGS. 5 (a) and 5 (b). The relative positions of the pixels (A, B) on both sides of the target boundary pixel (shown by diagonal lines) in FIG. In the example of the figure, A is (-1, -1),
B is (1,0). In the numbers, 1 pixel is 1.
The slope of the straight line passing through these two points is (-1-0) / (-1-
It is 1/2 in 1).

Therefore, as shown in FIG. 5 (b), the angle of inclination is cos ^-1 (1/2) = 30 degrees, and the angle of the normal line is 30 degrees + 90 degrees = 120 degrees. Therefore, it is determined that the direction of the object and the background is the vertical (Y) direction or the diagonal (Z2) direction.

FIG. 6 and FIG. 7 correspond to the invention described in claim 5 of the present invention, and the result of detecting the direction of the object and the background in the above-mentioned manner with respect to the object surrounding pixels (boundary pixels) of the object. Based on the object color and background color for the boundary pixel, and
It is a figure explaining the method of determining the object color after a color change, or the background color of a new background image. {See FIG. 6 (a)} Here, the case where the direction of the object and the background is the X direction will be described as an example. In the processing described here, the object color and the background color are detected from the original image {see FIG. 7 (b)}. At that time, the extracted object image {see FIG. 7 (c)} is used. For the object color after the color change, the extracted object image (color change image, see FIG. 7 (d)) with the changed color is used, and the new background color is the new background image {see FIG. 7 (e). } Is used.

First, as shown in FIG. 6A, the difference between the data of adjacent pixels is sequentially calculated from the target boundary pixel in the + direction of X. A pixel at a position where the difference value becomes smaller than a preset reference value is set as an object color or a background color. Similarly, the difference is calculated also in the-direction of X, and the pixel at a position smaller than the reference value is obtained and set as the background color or the object color.

Which of the detection pixels in the + -direction is the object color,
Whether to use the background color is based on the extracted object image. That is,
In the image of the extracted object {FIG. 7 (c)}, the fact that the background data (pixel value) is known in advance (usually black or white) is used.

The extracted object and the pixel data at the same position as the detected pixel on the background are examined, and it can be determined that the white or black position is the background color and the others are the object colors. The object color after the color change is the pixel data of the color-changed image at the same position as the position of the pixel determined to be the object color previously. Furthermore,
Similarly, the background color of the new background image is the pixel data of the background image at the same position as the position of the pixel previously determined to be the background color.

The calculation of the difference is performed in the same manner as in FIG.
Use each data of B and brightness. In the case of each RGB data, the position where the maximum value of the three data becomes smaller than the set reference value may be detected.

The standard value is, for example, about 4 to 8 gradations (2 to 3 bits) in the case of a normal 8-bit, 256-gradation image on a personal computer, workstation, etc., considering the influence of noise. Is considered good.

8 and 9 correspond to the invention according to claim 6 of the present invention, and the object color and the background color for the boundary pixel, the object color after the color change, or the background of the new background image. It is a figure explaining the 2nd method of determining a color. Similar to FIGS. 6 and 7, the case where the direction of the object and the background is the X direction will be described as an example.

The change in the data (pixel value) in the X direction is shown in FIG.
As in (a). Here, the slope line of the data change {the tangent line of the data change of FIG. 8 (a)} at the position of the target boundary pixel is obtained. The method of obtaining is to use data of pixels on both sides of a boundary pixel, or although not shown,
The change of the data is approximated by a cubic function, and the approximated 3
The maximum slope straight line may be obtained on the curve of the next function. Extend this line in the +-direction, and
As described above, the difference from the pixel data is sequentially calculated. The position where the calculation result is larger than the preset reference value is the pixel of the object color or the background color.

Which one is the object color and which is the background color may be the same as in the embodiment described with reference to FIGS. 6 and 7. Further, the method of obtaining the object color after the color change and the background color of the new background image are the same. Further, also regarding the method of calculating the difference, similarly, each RGB data or brightness is used. RG
When B data is used, noise-resistant processing can be realized by using the one having the largest slope (change in data) of the obtained straight line (R in the figure) as shown in FIG. 9C.

FIG. 10 corresponds to the invention described in claim 7 of the present invention, and determines the object color and background color for the boundary pixel, the object color after color change, or the background color of a new background image. It is a figure explaining the 3rd method to do. Similar to FIGS. 6, 7, 8 and 9, the case where the direction of the object and the background is the X direction will be described as an example.

The change in the data in the X direction is shown in FIG. Here, using the data of a plurality of pixels including the target boundary pixel, the equation of the curve representing the change of the data is obtained. For example, it is approximated by a cubic function. As a condition of curve approximation, the assumption that the value converges to a certain value in the-direction of the + direction of X is adopted. As a result, in the obtained curve, the pixel at the position where the convergence value is obtained is set as the pixel of the object color or the background color {see FIG. 10 (b)}. Which is the object color and which is the background color may be the same as in the embodiments of FIGS. 6, 7, 8 and 9. Further, the method of obtaining the object color after the color change and the background color of the new background image are the same.

Further, also regarding the method of calculating the difference, similarly, each RGB data or brightness is used. When using RGB data, noise-resistant processing can be realized by using the one having the largest difference in convergent values.

FIG. 11 is a diagram corresponding to the eighth aspect of the present invention and is a view for explaining the configuration of a color image color adjusting and image synthesizing apparatus. A color image is input by the image input device 1 such as an image scanner, and the image data is stored in the image holding memory 2. The object to be extracted is specified by the object extraction instruction means 3 such as a mouse. The object extraction unit 5 extracts the object. The extraction of the object is performed by, for example, an operator
The contour is extracted by designating an object or a background pixel by pixel, or the above-mentioned Japanese Patent Application No.
This is performed using the object contour detection method disclosed in Japanese Patent Publication No. 04-275593 “Object contour detection method, object extraction method, and image processing apparatus”.

These object extracting processes are controlled by the overall control unit 6. The extracted image is sent to the display memory 7 and displayed on the color display 8. The operator views the displayed image and checks the result. Further, the object data is sent to the image processing unit 9 which changes the color of the extracted object, changes the shape and the size, and performs various image processing.

The processed object image is sent to the image synthesizing unit 10 where the boundary portion is subjected to color mixing processing. In the color mixing processing, the direction of the object and the background described above, the detection of the object color and the background color, the calculation of the color mixing ratio K, the calculation of the new boundary pixel data, and the like are performed. As the image for synthesizing the object, the original image or the background image stored in the image storage means 11 of another image is used. The image of the composite processing result is displayed in the memory 7
Sent to and displayed on color display 8.

As described above, according to the present invention, an arbitrary object in an image is extracted, and further, the color, shape, size, etc. of the extracted image are changed so as to be superimposed on the original image or another background image. A method of adjusting the colors of a color image for combining (synthesizing) and creating a new image, and an image synthesizing method.
When an object image is obtained by designating an object to be extracted, the color of the extracted object is changed, and when superposed on the original image or another background image, with respect to each peripheral pixel of the extracted object, , The direction that is not along the contour of the extracted object, for example, the direction such as vertical, horizontal, diagonal, or the direction of the normal line to the boundary pixel around the object, and in the detected direction The difference between the data is calculated, and the pixel at the position where the difference value is smaller than the preset reference value is set as the object color or the background color, or the pixel data on both sides sandwiching the target boundary pixel. The straight line including is extended, the difference from the pixel data is sequentially calculated, and the pixel at the position where the calculation result is larger than the reference value is the object color,
Alternatively, the background color is used, or the data of a plurality of pixels including the target boundary pixel is used, and a curve representing the change in the data is obtained by an approximate function, and the pixel at the position where the convergence value is obtained in the obtained curve. Is characterized in that the pixel of the object color or the background color is used.

[0046]

As described above in detail, according to the color image color adjusting method, the image synthesizing method, and the apparatus of the present invention, in the case where the boundary area is wide (the color change is wide), By setting the data of the pixel at the position corresponding to it to the object color and the background color, the object color and the background color corresponding to the unclear boundary area can be detected. The effect is that highly accurate color mixing processing can be realized.

[Brief description of drawings]

FIG. 1 is a principle configuration diagram of the present invention (No. 1)

FIG. 2 is a principle configuration diagram of the present invention (No. 2)

FIG. 3 is a diagram showing an embodiment of the present invention (No. 1).

FIG. 4 is a diagram showing an embodiment of the present invention (part 2).

FIG. 5 is a diagram showing an embodiment of the present invention (part 3).

FIG. 6 is a diagram showing an embodiment of the present invention (No. 4).

FIG. 7 is a diagram showing an embodiment of the present invention (No. 5).

FIG. 8 is a view showing an embodiment of the present invention (No. 6).

FIG. 9 is a diagram showing an embodiment of the present invention (No. 7).

FIG. 10 is a view showing an embodiment of the present invention (No. 8).

FIG. 11 is a view showing an embodiment of the present invention (No. 9).

FIG. 12 is a diagram (part 1) illustrating a conventional method for detecting an object color and a background color.

FIG. 13 is a diagram (part 2) illustrating a conventional method for detecting an object color and a background color.

[Explanation of symbols]

1 Image input device 2 Image storage memory 3 Object extraction instruction unit (mouse) 5 Object extraction unit 6 Overall control unit 7 Display memory 8 Color display 10 Image synthesis unit 100 to 101, 101a, 102, 102a, 1020 to 1023, 1024 processing Step normal

Claims (8)

[Claims] 1. An arbitrary object in an image is extracted (100), a color of the extracted object is changed (101), and a peripheral portion of the color-changed object is mixed with a background image boundary in the image. (102) A color image color adjustment method, wherein the color-changed object is overlaid on the original background image (1).
As 02), with respect to the surrounding pixels of the extracted object, the direction of the object side and the background side is detected (1020), the object color corresponding to the surrounding pixels, the original background color, and the object after the color change. A color is determined (1021), and a data value of a new surrounding pixel is obtained using the object color and the background color, and the object color after the color change (1021).
2,1023) A color adjustment method for color images. 2. An arbitrary object in an image is extracted (100), the extracted object is subjected to predetermined image processing to change the color, shape, size, etc. (101a), and the image processed object is obtained. The peripheral part of the, is subjected to color mixing processing with an arbitrary background image boundary, and is superimposed (102a) is a color image image combining method, wherein the changed object subjected to the image processing is a method of superimposing it on an arbitrary background image. (102a), with respect to the surrounding pixels of the extracted object, the direction of the object side and the background side is detected (1020), the object color corresponding to the surrounding pixels, any background color, and the background of the background image. Determine the color (1024), the object color and background color,
An image synthesizing method for a color image, characterized in that data values of new surrounding pixels are obtained using the background color of the background image (1022, 1023). 3. A method of detecting a direction of an object side and a background side with respect to surrounding pixels of an extracted object according to claim 1, wherein a set of pixels separated from the surrounding pixels by a predetermined distance. A color adjustment method of a color image, in which a plurality of sets are obtained, the difference between the data of the obtained plurality of sets of pixels is calculated, and the direction of the set having the largest difference is the direction of the object and the background. Image composition method. 4. A method for detecting the directions of the object side and the background side with respect to the surrounding pixels of the extracted object according to claim 1 and 2, wherein: A color adjusting method for a color image and an image synthesizing method, characterized by detecting a normal line () and recognizing the detected normal direction as a direction of an object and a background. 5. The object color and background color corresponding to surrounding pixels according to claim 1, and the object color after color change,
Or, as a method of determining the background color of the background image, the detected object and the direction of the background, the difference between the data of the adjacent pixels of a plurality of pixels is calculated, the calculated difference value, from the preset value A pixel at a position that becomes smaller is detected, and the object color and the background color, and the object color after the color change, or the background color of the background image is determined based on the detected pixel. Color adjustment method and image composition method. 6. A method for determining the object color and background color corresponding to surrounding pixels, the object color after color change, or the background color of a background image according to claim 1 or 2, wherein: A straight line that passes through a plurality of pixels in the direction of the object and the background is detected, and a pixel at a position where the difference between the value determined by the straight line and the data value of the pixel in the object and the background direction is larger than a preset value is detected. Detected, based on the detected pixels,
A color image color adjusting method and an image synthesizing method, characterized in that an object color and a background color, and an object color after color change or a background color of a background image are determined. 7. A method for determining the object color and background color corresponding to surrounding pixels, the object color after color change, or the background color of a background image according to claim 1 or 2, wherein: From a pixel value of a pixel passing through a plurality of pixels in the direction of the object and the background, a curve that approximates the change in the pixel value is detected, and based on the pixel value of the pixel determined by the curve, the object color and the background color, and A color image color adjusting method and an image synthesizing method, characterized in that an object color after color change or a background color of a background image is determined. 8. A device for performing color adjustment of a color image and image composition, which is a memory means for storing input image data such as RGB signals.
(2) and specify any object in the input image,
Object extracting means for extracting an object from the specified result (3, 5)
An image processing means (9) for changing the color, shape, size, etc. of the extracted object, and superimposing the image processing result on the original image or another background image, and at the same time, the object and the background boundary. An image synthesizing device and a color synthesizing device (10) for performing color mixing processing.