JPH05236347A - Soft key generating device - Google Patents

Abstract

PURPOSE:To attain the synthesization of images of high quality by cutting a hard key graphic showing the area occupied by an object with a conical curved surface whose radius changes by the rotational angle for production of a soft key graphic, therefore eliminating the undesired swelling of a key area, attaining the change of soft quantity according to the places to eliminate the overs and shorts of soft quantity in each place, and furthermore dealing with even the moving blur of the object to be extracted. CONSTITUTION:An edge detecting part 2 and a motion vector detecting part 4 calculate the edge intensity and a motion vector on a hard key boundary line based on the inputted hard key signal and then produce a transformational parameter. A curved surface generating part 5 deforms a reference curved surface based on the transformational parameter to produce a transformational curved surface. At the same time, a soft key generating part 6 cuts the boundary line part of a hard key graphic corresponding to the hard key signal based on the transformational curved surface to produce a soft key graphic.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention extracts an object area having a specific meaning, such as a person or a flower, from a normal image, arranges these at appropriate positions, and creates a composite image using key information. , A softkey generation device for generating softkey information from a hardkey figure showing an area occupied by an object.

SUMMARY OF THE INVENTION The present invention extracts an object area having a specific meaning such as a person or a flower from a normal image, arranges these at appropriate positions, and sets key information (information indicating an area occupied by the object). When creating a composite image using), a hard key figure showing the area occupied by the object has a conical curved surface whose radius changes according to the rotation angle (depending on the edge strength and motion vector of the extracted image on the hard key boundary line). By creating a soft key figure by scraping the hard key figure with a conical curved surface, the unnecessary bulge of the key area is eliminated and the amount of software can be changed depending on the location, eliminating excess or deficiency depending on the location. Furthermore, a soft key figure that can correspond to the motion blur of the extracted object is generated, and high quality image synthesis is performed.

[0003]

2. Description of the Related Art Generally, in synthesizing images, a region having a specific meaning such as a person or a flower is cut out from a normal image, and these are arranged at appropriate positions.
A composite video is created based on the key information generated when these images are cut out.

In this case, when extracting a specific object area manually or semi-automatically from a natural image, the key information (area information) obtained is a hard key of "0" or "1". It is necessary to generate softkey information from the key information.

FIG. 10 is a block diagram showing a main part of a conventional soft key generation device for performing such conversion processing.

The softkey generation device shown in this figure is provided with a low-pass filter circuit 101. When a hardkey signal as shown in FIG. 11A is input, this hardkey signal is lowpassed to obtain the softkey signal shown in FIG. The soft key signal shown in b) is converted to blur the periphery of the key area.

Then, as shown in FIG. 12, a plurality of frames of pixel data (pixel data) arranged in a two-dimensional manner are provided to hold the video content.

[0008]

However, the above-mentioned conventional soft key generation device has the following problems.

First, since the key area swells due to the low-pass processing, an extra background area is also extracted when the image of the object is extracted from the natural image as shown in FIGS. 11 (a) and 11 (b).

Further, since the amount of software is constant over the entire screen, the amount of software may be excessive or insufficient depending on the location.

Further, since the softening is performed uniformly in all directions, it is impossible to deal with the motion blur of the object.

Therefore, in the conventional soft key generation device, when obtaining a high quality image, it is necessary to manually process the image one frame at a time to generate the soft key information, which is very troublesome. There was a problem that it would end up.

In view of the above circumstances, the present invention provides a hard key figure showing an area occupied by an object with a conical curved surface whose radius changes according to the rotation angle (the edge strength and the motion vector of the extracted image on the hard key boundary line are Dependent conical curved surface) allows hard key figures to be shaved to create soft key figures, eliminating unnecessary bulging of the key area and changing the amount of software depending on the location. It is an object of the present invention to provide a softkey generation device capable of eliminating excess or deficiency and generating a softkey figure capable of coping with the motion blur of an object to be extracted, and performing high-quality image synthesis. ..

[0014]

In order to achieve the above object, a soft key generation apparatus according to the present invention is an apparatus for combining a plurality of images to synthesize a new image, based on an input hard key signal. A curved surface deformation parameter detection unit that calculates edge strength and motion vector on a hard key boundary line to create a deformation parameter, and a deformation surface obtained by deforming the reference curved surface by the deformation parameter obtained by this curved surface deformation parameter detection unit A curved surface generating section; and a hard key deforming section for creating a soft key figure by shaving the boundary line portion of the hard key figure corresponding to the hard key signal based on the deformed curved surface obtained by the curved surface generating section. Is characterized by.

[0015]

In the above configuration, the curved surface deformation parameter detection unit calculates the edge strength and the motion vector on the hard key boundary based on the input hard key signal to create a deformation parameter, and the curved surface deformation unit generates the curved surface deformation parameter. A reference curved surface is deformed by the deformation parameter obtained by the detection unit to generate a deformed curved surface, and a hard key figure corresponding to the hard key signal is generated by the hard key deformation unit based on the deformed curved surface obtained by the curved surface generation unit. The boundary part of is deleted and a softkey figure is created.

[0016]

FIG. 1 is a block diagram showing an embodiment of a soft key generation device according to the present invention.

The softkey generation device shown in this figure includes a boundary detection unit 1, an edge detection unit 2, a video delay unit 3, a motion vector detection unit 4, a curved surface generation unit 5, and a softkey generation unit 6. It is equipped with a conical shape that depends on the detection result by detecting the edge strength and motion vector of the extracted image on the hard key boundary line indicated by the hard key signal according to the characteristics of the image to be extracted as the video signal. A curved surface is generated and a soft key signal is generated from the hard key signal.

The boundary detecting section 1 is a circuit for detecting a boundary of an image to be extracted which is indicated by the hard key signal when the hard key signal is input, and is indicated by the hard key signal as shown in FIG. Among the pixels to be processed, the output pixel y for the pixel x is determined based on the pixel to be processed, for example, the pixels a and c above and below the pixel x in the center, and the pixels b and d on the left and right, and The boundary signal is supplied to the edge detection unit 2, the motion vector unit 4, and the curved surface generation unit 5.

In this case, if the condition that the pixel x is "0" and any one of the pixels a, b, c, d is "1" is satisfied, the output pixel y becomes "1", If the above condition is not satisfied, the output pixel y becomes "0".

As a result, when the hard key signal shown in FIG. 3 (a) is supplied corresponding to the video signal shown in FIG. 3 (c), the boundary detection unit 1 outputs the boundary signal shown in FIG. 3 (b). It is generated and supplied as a boundary signal to the edge detection unit 2, the motion vector unit 4, and the curved surface generation unit 5.

The edge detection unit 2 extracts each pixel in the edge region from the video signal based on the boundary signal output from the boundary detection unit 1 and generates an edge signal indicating the edge strength from each pixel. As shown in FIG. 4, each pixel in the edge region is extracted from the video signal based on the boundary signal output from the boundary detection unit 1 to generate an edge signal indicating the edge strength, and the edge signal is generated by the curved surface generation unit 5. Supply.

In this case, as an edge detecting method, various methods, for example, a pixel to be processed among the pixels extracted from the video signal, for example, the pixel x in the center and the pixels a above and below the pixel x, c, the intensity E of the output pixel y is determined by a method of detecting the maximum absolute difference between the left and right pixels b and d, and determining the intensity E of the output pixel y corresponding to the pixel x. Then, this is supplied to the curved surface generator 5 as an edge signal.

The video delay unit 3 is a circuit that delays the video signal by one frame when the video signal is supplied, and the motion vector detection unit detects the video signal of the previous frame obtained by the delay operation. Supply to 4.

The motion vector detecting section 4 extracts each pixel of the boundary portion from the video signal of the current frame on the basis of the boundary signal output from the boundary detecting section 1, and as shown in FIG. The motion vector of the edge portion is detected based on the pixel and each pixel of the previous frame to generate a motion vector signal, which is supplied to the curved surface generation unit 5.

In this case, the elements u and v of the motion vector that minimize the total sum expressed by the arithmetic expression shown in the upper part of FIG.
After that, the magnitude M of this motion vector and the direction α are calculated by the arithmetic expression shown in the lower part of FIG.

The curved surface generator 5 takes in the edge signal output from the edge detector 2 and the motion vector signal output from the motion vector detector 4 based on the boundary signal output from the boundary detector 1. At the same time, the edge signal and the motion vector signal are used as parameters to generate a curved surface signal indicating a conical curved surface depending on the edge strength and the motion vector of the extracted image on the hard key boundary line, and this is supplied to the soft key generation unit 6. ..

In this case, as shown in FIG. 6, the shape of the curved surface becomes a conical shape whose upper portion is widened as the edge strength is smaller.

Then, in the portion of "Z = 1.0" of the reference phase, a circle having a sectional shape of the formula shown in FIG. 7 is formed, and the smaller the edge strength is, the larger the sectional shape becomes, and the conical shape widens upward. Become.

Further, as the magnitude M of the motion vector is larger, the elliptic cone having the major axis a and the minor axis b based on the arithmetic expression shown in FIG. 8 and rotated according to the direction α of the motion vector as shown in FIG. Take shape. The characters "b1", "b2", and "a1" used in the arithmetic expression shown in FIG. 8 are characters indicating constants.

The soft key generator 6 is the curved surface generator 5
Is a circuit that generates a soft key signal from the hard key signal based on a curved surface signal output from the curved surface generation unit 5, which holds the hard key signal in frame units and is output from the curved surface generation unit 5 as shown in FIG. Based on the curved surface signal, the periphery of the hard key figure indicated by the hard key signal is cut off to generate a soft key figure, and this is output as a soft key signal.

In this case, when the apex of the conical shape makes a round on the boundary line, the periphery of the hard key figure is shaved by a soft amount according to the shape of the conical shape, that is, the edge strength and the change of the motion vector. Is created.

As described above, in this embodiment, the soft key figure is created by cutting the periphery of the hard key area with the conical curved surface corresponding to the edge strength and the motion vector of the boundary portion of the hard key figure. , A softkey figure can be created by scraping a hardkey figure showing the area occupied by an object with a conical curved surface (conical curved surface depending on the edge strength and motion vector of the extracted image on the hardkey boundary line). The unnecessary bulge of the key area can be eliminated by it, and the soft amount can be changed depending on the location to eliminate the excess or deficiency depending on the location, and the soft key figure that can also correspond to the motion blur of the extracted object Can be generated, and high-quality video composition can be performed.

Further, in the above-mentioned embodiments, the present invention has been described by taking a monochrome image as an example, but the present invention may be applied to a color image. In this case, the color difference between adjacent pixels may be used as the edge information.

In addition, when it is desired to manually control the amount of software in the production, a curved surface is generated by manually giving the parameters of the curved surface signal on the boundary line only in the first frame, and the boundary is generated in the subsequent frames. With respect to the pixels on the line, the entire surface of the frame can be automatically processed by cutting the hard key figure using the curved surface of the pixel on the boundary line of the corresponding previous frame.

[0035]

As described above, according to the present invention, the hard key figure showing the area occupied by the object has a conical curved surface whose radius changes according to the rotation angle (the edge strength of the extracted image on the hard key boundary line). , A conical curved surface that depends on the motion vector) can create a softkey figure by scraping the hardkey figure, eliminating unnecessary bulging of the key area and changing the amount of software depending on the location. Thus, it is possible to eliminate excess or deficiency depending on the place, and to generate a softkey figure that can cope with the motion blur of the object to be extracted, and to perform high-quality image synthesis.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a soft key generation device according to the present invention.

FIG. 2 is a schematic diagram showing an example of a boundary detection operation of a boundary detection unit shown in FIG.

FIG. 3 is a schematic diagram showing an example of a boundary detection operation of a boundary detection unit shown in FIG.

FIG. 4 is a schematic diagram showing an example of an edge strength detecting operation of the edge detecting section shown in FIG.

5 is a schematic diagram showing an example of motion detection operation of the motion vector detection unit shown in FIG.

FIG. 6 is a schematic diagram showing an example of a curved surface creation operation of the curved surface generation unit shown in FIG.

FIG. 7 is a schematic diagram showing an example of a curved surface creation operation of the curved surface generation unit shown in FIG.

8 is a schematic diagram showing an example of a curved surface creation operation of the curved surface generation unit shown in FIG.

9 is a schematic diagram showing an example of a softkey figure creation operation of the softkey generation unit shown in FIG.

FIG. 10 is a block diagram showing a main part of a conventionally known soft key generation device.

11 is a waveform chart showing an operation example of the low-pass filter shown in FIG.

FIG. 12 is a schematic diagram showing a configuration example of a pixel processed by the softkey generation device shown in FIG.

[Explanation of symbols]

 1 boundary detection unit 2 edge detection unit (curved surface deformation parameter detection unit) 3 video delay unit 4 motion vector detection unit (curved surface deformation parameter detection unit) 5 curved surface generation unit 6 soft key generation unit (hard key deformation unit)

Claims (1)

[Claims] 1. A device for synthesizing a new image by combining a plurality of images, curved surface deformation for generating a deformation parameter by calculating an edge strength and a motion vector on a hard key boundary line based on an input hard key signal. A parameter detection unit, a curved surface generation unit that deforms the reference curved surface by the deformation parameter obtained by the curved surface deformation parameter detection unit to generate a deformed curved surface, and the hard key based on the deformed curved surface obtained by the curved surface generation unit. A soft key generation device comprising: a hard key deforming unit that creates a soft key graphic by cutting off a boundary line portion of the hard key graphic corresponding to a signal.