JPH0338780A - Method and device for synthesizing image - Google Patents

Abstract

PURPOSE:To obtain a natural image by performing a ratio mixing arithmetic operation after changing the value of a slip-in image outside the masking area to that of the slip-in image inside the masking area. CONSTITUTION:A mask image smoothing means 1 is provided which performs a smoothing processing on a mask image to instruct an area where each image should be positioned when a background image is synthesized with the slip-in image. Also, an image value changing means 2 is provided which performs a change processing for the image value of the slip-in image within a prescribed range outside the mask image with the value of the picture element of the slip-in image inside the mask image. A synthetic image can be generated by performing the ratio mixing arithmetic operation after changing the value of the slip-in image outside the masking area to that of the slip-in image inside the masking area. Thereby, the natural image can be obtained without affecting on the synthetic image even when the color of the slip-in image outside the masking area is remarkably different. Also, image information in the neighborhood of a boundary can be prevented from being missed.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an image synthesis method and apparatus for synthesizing a plurality of images consisting of pixels of light and shade by using a mask image.

(Prior Art) The present applicant previously filed an application for the invention of an image synthesis device invented by the present inventor (Japanese Patent Application No. 83-258375).

The image compositing device of this prior application smoothes a mask image that specifies the area to be inlaid when compositing one image with another image, and calculates the ratio mixing of the two images using the values of the smoothed mask image. By doing so, it is possible to prevent the appearance of discomfort due to sudden density differences at the boundaries, and has the advantage that images that do not cause discomfort can be synthesized at high speed with a simple configuration. However, I cut out a person from an image of a person standing against the blue sea,
When compositing a person onto an image of white sand, the blue ocean in the background will be included in the area where ratio mixing calculations are performed.
This has the disadvantage that a blue boundary appears between the white sand and the person in the composite image, resulting in an unnatural composite image. Another option is to reduce the mask before compositing so that the blue sea does not enter the area where the ratio mixing calculation is performed, but this method has the disadvantage that the area around the boundary of the person is missing, resulting in an unnatural image. Ta.

(Problems to be Solved by the Invention) The present invention improves the invention of the earlier application, and when smoothing a mask image and performing a ratio mixing operation so as not to cause a density step at the boundary, the excess area becomes a composite image. To provide an image compositing method and device capable of compositing images with high quality and without any discomfort, without entering the inside of the image or missing the boundaries of the fitted image.

(Means for Solving the Problems) The image synthesis method of the present invention smoothes a mask image indicating the area where the background image and the inset image should be, and uses the values of the smoothed mask image to create the area between the background image and the inset image. When performing the ratio mixing calculation, the pixel value of the fitted image outside the mask image is changed to the pixel value of the fitted image inside the mask image, and then the ratio mixing calculation is performed.

Further, in one aspect of the image synthesis apparatus for carrying out the above-mentioned method of the present invention, as shown in FIG. a mask image smoothing means 1 that performs smoothing processing on the mask image; and a pixel value change unit that performs processing to change the pixel values of the inset image in a predetermined range outside the mask image by the pixel values of the inset image that are inside the mask image. and a mixing operator stage 3 that performs a ratio mixing operation between the background image and the inset image changed by the pixel value changing means 2, using the values of the smoothed mask image.

(Function) According to the image synthesis method of the present invention, the value of the inset image outside the mask area is changed to the value of the inset image inside the mask area, and then a ratio mixing operation is performed to generate a composite image. 2. Even if the colors outside the mask area of the inset image are extremely different colors, a natural image can be obtained without affecting the composite image. Also, image information near the border may be lost.)

Further, the image synthesis device of the present invention performs ratio mixing of the background image, the inset image, and the image by smoothing a predetermined range outside the mask image by the pixel value changing means 2, for example, by the mask image smoothing means 1. The pixel values of the embedded image in the range (0<K<1) are changed based on the pixel values of the embedded image inside the mask image. Images outside the mask area of the inset image are excluded from calculation targets when ratio mixing is performed by the mixing calculation means 3, so as mentioned above, the color outside the mask area of the inset image is extremely different from the inner color. Even if the colors are different, a natural image can be obtained without affecting the composite image.

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

FIG. 2 is a block diagram of this embodiment, and the image synthesis apparatus of this embodiment includes a background image memory 11 for storing a background image, an inset image memory 12 for storing an inset image, and an image that is the same as the background image and the inset image. A mask image memory 13 that stores a set mask image consisting of a size, a mask image smoothing device 14 that smoothes the mask image, and an inset when the values of the mask image and the smoothed mask image meet specific conditions. A pixel value changing device 15 that changes the pixel value of an image by 11, a mixing calculation device 16 that performs a ratio mixing calculation of the background image and the inset image to generate a composite image, a composite image memory 17 that stores the composite image, and an output display bag [ 1
It has 8.

The operation of the image synthesis apparatus of this embodiment is shown in the operation flow diagram of FIG.

When combining the background image and the inset image, a mask image is generated that indicates in which area of the combined image these images should be placed, and is input into the mask image memory 13 (see FIG. 3). FIG. 4 is an explanatory diagram showing a part of the mask image on the actual flow side, and one grid represents one pixel. In the figure, pixel value 1 is inset during synthesis ii! In the area where the it image is displayed, the pixel value O is the area where the background image is displayed during synthesis, and the thick line is the boundary between the two images.

The mask image smoothing device 14 smoothes the generated mask image (FIG. 3). For example, 3 shown in FIG.
When the mask image shown in FIG. 4 is smoothed using a ×3 smoothing filter, a smoothing mask 1iIiii &
is generated.

The mixing arithmetic unit 15 first detects an area outside the mask image (the area with the pixel value O in FIG. 4) and where the value K of the smoothed mask image satisfies O<K<1, that is, the diagonal line in FIG. For the area indicated by , the pixel values of the inset image are replaced with the pixel values of the area inside the mask image (Fig. 3 -). Specifically, the pixel A in FIG. 8 is replaced with the pixel value of AI located in the inner area of the mask image among the four pixels A1 to A4 that are vertically and horizontally adjacent to each other. If two pixels (B, ,B2) out of four adjacent pixels (B, ~B,) are inside the mask image, as in B,
Replace with the average value of these pixels (Bl+82)/2. In addition, as in C, if none of the four pixels adjacent to each other in the upper, lower, left, and right directions are inside the mask image, among the four pixels C3 to C4 adjacent in the diagonal direction, C3, which is inside the mask image, Replace with pixel value. In this way, for the area shown by diagonal lines in Fig. 7, first, the adjacent 4
If there is a pixel within the mask area among the two pixels, it is replaced with this pixel value. At this time, if a plurality of pixels are within the mask area, their average value is substituted. If none of the four vertically adjacent pixels are within the mask area, the four diagonally adjacent pixels are replaced in the same way.

Next, the following ratio mixing operation is performed on the background image and the inset image replaced as described above using the value K of the smoothed mask image to obtain a composite rifi image. (Figure 3■
).

Vo= (1-K)*V+++to*V+t-(1)■o
Pixel value of the two composite images ■,: Pixel value of the background image v1 □: Pixel value of the replaced inset image: Pixel value of the smoothed mask image The composite image generated in this way is stored in the composite image memory. 17, and the operator views the composite image on the output display device 18. If the operator is satisfied with the composite image, the process is terminated, and if unsatisfactory, the operator may change the smoothing characteristics of the mask image smoothing device, etc. and re-execute the process (Fig. 3 ■)
.

In this example, a 3x3 filter with a uniform kernel as shown in Fig. 5 was used to smooth the mask image, but a filter with a weighted center part or a filter of a different size such as 5x5 may also be used. No problem.

Also, change the pixel value of the inset image by lli! The replacement method does not necessarily have to be as in this embodiment, and may be replaced with the average value of the 8 pixels in the upper, lower, left, right, and diagonal directions that are within the mask.

Furthermore, in this embodiment, it has been explained that the pixel values of the inset image are replaced and then the ratio mixing calculation is performed, but as shown in FIG. A similar composite image can also be generated by determining whether it is outside the mask area and performing a ratio mixing operation by using the value inside the mask area as the value of the inset image. Note that this method has the advantage that a composite image can be generated without the need to rewrite the contents of the inset image memory because ratio mixing calculations are performed sequentially.

(Effects of the Invention) As explained above, according to the present invention, the ratio mixing operation is performed while changing the value of the inset image outside the mask area to the value of the inset image inside the mask area, and a composite image is created. Therefore, even if the colors outside the mask area of the inset image are extremely different colors, a natural image can be obtained without affecting the composite image. Furthermore, image information near the boundary will not be lost.

[Brief explanation of drawings]

Figure 1 is a block diagram showing the basic configuration of the present invention, Figure 2 is a block diagram showing the basic configuration of the present invention.
The figure is a block diagram of an image synthesis device according to an embodiment of the present invention.
FIG. 3 is a flowchart explaining the operation in this embodiment, FIG. 4 is a diagram showing a part of the mask image in this embodiment, and FIG. 5 is an example of a smoothing filter used in this embodiment. Figure 6 shows a smoothed mask image obtained by applying the filter shown in Figure 5 to the mask image shown in Figure 4, and Figure 7 shows how the pixel values of the inset image are changed when performing ratio mixing calculations. FIG. 8 is a diagram showing the regions, and FIG. 8 is a diagram showing which pixel values are to be changed to the pixel values of the embedded image. FIG. 9 is an operation flow diagram in another embodiment. l...Mask image smoothing means, 2...Pixel value changing means, 3...Mixing calculation means, 11...Background image memory, 12...Inset image memory, 13...Mask image memory , 14: mask image smoothing device, 15: pixel value changing device, 16: mixing operation device, 17: composite image memory, 18: output display device. +1- Figure 3 Figure 4

Claims (2)

[Claims] (1) Smooth the mask image that indicates the area where the background image and the inset image should be, and use the values of the smoothed mask image to perform a ratio mixing operation between the background image and the inset image. An image synthesis method characterized by changing pixel values of an image to pixel values of an inset image inside a mask image and then performing a ratio mixing operation. (2) Mask image smoothing means for smoothing the mask image, and pixel value changing for changing the pixel values of the inset image in a predetermined range outside the mask image by the pixel values of the inset image inside the mask image. and mixing calculation means for performing a ratio mixing calculation of a background image and an inset image changed by the pixel value changing means, using the values of the smoothed mask image. Device.