JPH089377A - Image signal encoder and image signal decoder - Google Patents

Abstract

PURPOSE:To provide an encoding effect equivalent to the one in the case of performing encoding without performing division by enabling the search of a motion vector over a division boundary and minimizing the addition of encoding data accompanying the overlap of areas at the time of dividing picture signals and performing the encoding. CONSTITUTION:In an image division means 10, for input digital image signals, the image signals of a present frame are divided into some pieces of image data. A divided image magnification means 11 adds dummy data in the horizontal and vertical directions of the encoded object frames of the respective divided images and magnifies an image size at the time of encoding the divided images. In a frame memory 13, the image signals of the frames in the past or future for reference in the encoding are stored. At the time of reading the image signals from the memory 13, a part or all of the image signals of the frames in the past or future of the divided images other than the encoding object divided image is added and a reference image size is magnified. The encoding is performed after detecting a vector in a motion search and encoding means 21 and output from a multiplexing means 14 is performed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a high-efficiency coding of a television signal or the like, and a motion search area in the case where one image signal is divided into a plurality of pieces for encoding and the divided images are superposed on each other. This will improve the image quality and
The present invention relates to an image signal encoding device and an image signal decoding device that can minimize the amount of encoded data for overlapping image areas.

[0002]

2. Description of the Related Art When one image signal is divided into a plurality of pieces for encoding, each divided image is regarded as one image and encoded. In the case of division and encoding, parallel processing is possible, and real-time encoding of high-resolution video such as HDTV can be easily realized. Also, it becomes possible to use a coding tool for standard TV.

FIG. 8 is a diagram for explaining an example of a conventional division coding method, and FIG. 9 is a diagram showing an example of division of an image signal. For example, as shown in FIG. 9, one image signal is divided into four, and each divided screen is A, B, C, D,
Assume that motion compensation is used during the image coding scheme.

First, the image input from the input terminal 901 is divided into four divided image signals by the image signal dividing section 902. The divided image signal 903 is transmitted to each of the encoding devices 904-9.
Encoding is performed at 07, and the output 908 of each encoding device is multiplexed by the multiplexing device (MUX) 909 and output to the output terminal 910 as encoded data of the input image.

According to such a method, each divided image can be individually encoded and decoded. As a result, when encoding is performed with N divisions, it is possible to perform encoding in about 1 / N the time of encoding without division.

[0006]

[Problems to be Solved by the Invention]
Consider the prediction near the divided boundaries. FIG. 10 is a diagram for explaining the cause of the reduction in coding efficiency during division. FIG. 10 (A) shows that when the prediction is performed without dividing the image signal,
FIG. 10B shows a case where the prediction is performed after the division. Figure 1
When b is predicted at 0 (A), b'of the reference image can be predicted as a prediction signal, but FIG.
, B and b'are different divided image signals, so b '
Cannot be used as a prediction signal and must be predicted within the same divided image signal. Therefore, when the image signal is divided and encoded, there is a problem that the encoding efficiency decreases due to an increase in prediction error.

Further, when encoding a high-resolution image signal such as HDTV, if the motion search range is not limited to within the divided image with the divided image as the standard TV size, the motion search range of the reference image frame is expanded. Because the image size of the divided image frame to be encoded differs from that of the reference image frame, the normal encoding tool cannot be used.

Further, if the divided images are made to overlap each other, the search range of the prediction signal is widened and the motion compensation efficiency is improved, but the overlapped portion is double coded, and the coding efficiency is increased. There is a problem in that

[0009]

The present invention has been made in order to solve the above problems, and is provided with each means as shown in FIG.

FIG. 1A shows an example of the configuration of an image signal coding apparatus corresponding to the invention described in claim 1. Image dividing means 1
0 is a means for dividing the image signal of the current frame into some image data with respect to the input digital image signal.
The divided image enlarging means 11 is means for enlarging the image size by adding dummy data in the horizontal and vertical directions of the encoding target frame of each divided image when encoding the divided image. The frame memory 13 stores image signals of past or future frames for reference in encoding. When reading the image signal from the frame memory 13 for reference, the image signal of the past or future frame of the divided image to be encoded is added to the image signal of the past or future frame of the divided image other than the divided image to be encoded. Some or all of them are added and the size of the referenced image is enlarged. The motion search / encoding means 12 is the divided image enlarging means 1
For the divided image of the current frame enlarged by 1,
It is a means for performing coding after detecting a motion vector using an enlarged divided image of a past or future frame. The multiplexing means 14 multiplexes the encoded data of each divided image and outputs it.

FIG. 1B shows an example of the configuration of an image signal coding apparatus corresponding to the invention described in claim 2. Image dividing means 2
0 is a means for dividing the image signal of the current frame into some image data with respect to the input digital image signal.
The frame memory 22 stores image signals of past or future frames for reference in encoding. When reading an image signal from the frame memory 22 for reference, the image signal of the past or future frame of the divided image other than the encoding target divided image is added to the image signal of the past or future frame of the divided image to be encoded. Some or all of them are added and the size of the referenced image is enlarged. The motion search / encoding means 21 applies to the divided image of the current frame,
It is a means for performing coding after detecting a motion vector using an enlarged divided image of a past or future frame. The pseudo data adding means 23 is means for adding coded data for expanding the image size to the coded data of each divided image. The multiplexing means 24 multiplexes the encoded data of each divided image and outputs it.

FIG. 1C shows an example of the configuration of the image signal decoding apparatus according to the third aspect. This image signal decoding apparatus is shown in FIG.
It is a device for decoding coded data coded by the image signal coding device shown in (A) or (B).

The demultiplexing means 30 separates the received encoded data for each divided image. The decoding means 31 is means for decoding the encoded data of the separated divided images.
The frame memory 34 stores image signals of past or future frames for reference in decoding. When reading an image signal from the frame memory 34 for reference, a part of the image signal of the past or future frame of the divided image other than the decoding target divided image is added to the image signal of the past or future frame of the decoding target divided image. Or all of them are added and the image size to be referred to is enlarged. The divided image reproducing means 32 is means for reproducing the divided image of the current frame including the enlarged area from the decoded divided image information by using the divided image of the enlarged past or future frame. The enlarged area deleting means 33 is means for deleting the enlarged area in the divided image from the reproduced divided image. The image synthesizing means 35 synthesizes and outputs the reproduced divided images.

[0014]

In the image signal coding apparatus of the present invention, the image signal is first divided into N pieces, and the divided image signal is encoded. At this time, the reference frame used for prediction as shown in FIG. 2 is not only the past or future frame of the encoding target divided image, but also a part or all of the past or future frame of the other divided image, The area that can be referred to as the prediction signal is expanded and the same area as that when encoded without division can be referred to.

Since the size of the reference image frame is larger than that of the divided image frame to be encoded due to the enlargement of the referenceable area, the apparent image size of the divided image frame to be encoded is enlarged. To do.

In each of the divided images, the processing of the area where the divided images which are generated by the enlargement of the image size overlap each other (hereinafter, referred to as an enlarged area) uses one of the following methods as shown in FIG.

(1) As shown in FIG. 3A, the enlarged area is encoded by adding a dummy image signal (for example, an area in which all pixel values are 0). The dummy image signal used here is determined so that the size of the data generated when encoded is as small as possible.

(2) As shown in FIG. 3 (B), the enlarged area is not encoded, and dummy encoded data prepared in advance is output. The dummy encoded data used here is
It is the encoded data necessary to represent the enlarged area.

FIG. 4 is a diagram showing a search example of a motion vector in the enlarged area. When past or future images required for motion search or motion compensation are decoded by the local decoder, the decoded signal in the enlarged area is not written in the frame memory. Further, when performing motion search or motion compensation using the enlarged region as a prediction signal, a search is performed using a decoded image in a past or future frame memory for a divided image other than the encoding target divided image as a reference image.

Next, the image signal decoding apparatus will be described.
In the image signal decoding device, the coded data divided into N pieces are decoded in the decoding device for each divided image. At the time of decoding, the image data obtained by decoding the encoded data of the enlarged area is not written in the frame memory. Further, when the motion vector indicates the enlarged region as the prediction signal, similarly to the encoding side, the divided image other than the already decoded past or future encoding target divided image is used as the prediction signal.

According to the above means, the motion vector search area is not limited to the divided image. In addition, the expansion area added to the divided image due to the expansion of the motion vector search range is not decoded.

[0022]

EXAMPLES Examples of the present invention will be described below. In the present embodiment, it is assumed that the HDTV image signal is divided into 4, and a standard TV encoding device and a decoding device using MPEG as an image encoding method are used.

First, the image signal coding apparatus will be described. FIG. 5 shows an example of the configuration of the image signal coding apparatus according to claim 1 of the present invention. First, the HDTV image input from the input terminal 101 is divided into four standard TV size divided image signals 103 by the image signal dividing unit 102. The divided image signal 103 is input to each of the divided image enlarging units 104 to 107, and the enlarged area filled with the pixel value 128 is added to enlarge the size of the divided image. The enlarged divided images are encoded by the encoding units 108 to 111.

The motion search units 112 to 115 of each coding apparatus
In, the motion vector is obtained by using a part or all of the shared frame memory 116 (HDTV size) as the reference image 117. In the local decoder output image 118 of each encoding device, the enlarged area deletion unit 119 deletes the enlarged area portion added to each divided image, restores the standard TV size, and performs motion compensation prediction of the next frame. It is stored in the shared frame memory 116. The coded data 120 of the divided image is multiplexed by a multiplexer (MUX) 121, and is output terminal 12 as coded data of the input image.
2 is output.

FIG. 6 shows an example of the configuration of an image signal coding apparatus according to claim 2 of the present invention. First, the HDTV image input from the input terminal 201 is converted by the image signal dividing unit 202 into 4
It is divided into a plurality of standard TV size divided image signals 203. The divided image signal 203 is generated by the encoding unit 2 of each encoding device.
Encoding is performed from 04 to 207.

Motion search units 208 to 211 of the respective coding devices
In, the motion vector is obtained by using a part or all of the shared frame memory 212 (HDTV size) as the reference image 213. The local decoder output image 214 of each encoding device is stored in the frame memory 212 for motion compensation prediction of the next frame. Each encoding unit 204
The encoded data of 207 to 207 are stored in the pseudo data adding units 215.
At ~ 218, dummy macroblocks necessary to represent the enlarged area are added, the image size is enlarged,
The divided image is encoded data 219. The coded data 219 of the divided images are multiplexed by the MUX 220 and output to the output terminal 221 as the coded data of the input image.

FIG. 7 shows an embodiment of the image signal decoding apparatus of the present invention. First, multiplexed encoded data is input from the input terminal 301, and the demultiplexer (DMUX) 3
At 02, it is divided into four pieces of encoded data 303 of the divided image signal. The encoded data 303 of the divided image is transferred to each decoding unit 3
Decrypt at 04 to 307. The frame memory 320 (HDT) shared with the information decoded by each of the decoding units 304 to 307
The divided images are reproduced by the image reproducing units 312 to 315 based on the V size image. At this time, the enlarged area determination unit 308
311 determines whether the decoded information to be reproduced is the information of the enlarged area or the information in the divided image, and if it is the information of the enlarged area, the switches 316 to 319 are turned off to determine whether the information is in the divided image. Turn it on. Switches 316-319
The enlarged area is not reproduced by switching the
The V-size divided image is reproduced. Regenerated divided image 3
The divided image synthesizing unit 323 synthesizes the image 22 into an HDTV size image and outputs it to the output terminal 324.

In the above-described embodiment, the input image is HDT.
Although the V size, the divided image is the standard TV size, and the division is 4, it is not limited to this, of course. Further, the pixel value for image enlargement in the embodiment of the invention described in claim 1 is arbitrary. Further, the coded data added to represent the enlarged area in the embodiment of the invention described in claim 2 can be arbitrarily set as long as it matches the format of the image coding method used, but the amount of information to be added Is preferably less.

[0029]

As described above, according to the present invention,
When the image signal is divided and encoded, it is possible to search for a motion vector that crosses the division boundary. Further, it is possible to minimize the increase in coded data due to the overlapping of areas. As a result, it is possible to realize the same coding efficiency as that when the coding is performed without division by using a plurality of coding means that can process only a size smaller than the image to be coded.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is a diagram showing an example of expanding a search range of a motion vector.

FIG. 3 is a diagram showing a processing example of an enlarged area at the time of encoding.

FIG. 4 is a diagram showing a search example of a motion vector in an enlarged area.

FIG. 5 is a block configuration diagram of an embodiment of the image signal encoding device according to claim 1;

FIG. 6 is a block configuration diagram of an embodiment of an image signal encoding device according to claim 2;

FIG. 7 is a block configuration diagram of an embodiment of an image signal decoding apparatus of the present invention.

FIG. 8 is a diagram showing a configuration example of a device for dividing and encoding a conventional image signal.

FIG. 9 is a diagram showing an example of division of an image signal.

[Fig. 10] Fig. 10 is a diagram for describing the cause of a decrease in coding efficiency during division.

[Explanation of symbols]

 10 image dividing means 11 divided image enlarging means 12 motion search / encoding means 13 frame memory 14 multiplexing means 20 image dividing means 21 motion search / encoding means 22 frame memory 23 pseudo data adding means 24 multiplexing means 30 demultiplexing means 31 Decoding means 32 Divided image reproducing means 33 Enlarged area deleting means 34 Frame memory 35 Image combining means

Claims (3)

[Claims] 1. An image signal encoding device for dividing a digital image signal of an image signal of a current frame into some image data and encoding each divided image signal,
When encoding a divided image, means for increasing image size by adding dummy data in the horizontal and vertical directions of the encoding target frame of each divided image, and image signal of past or future frame of the encoding target divided image To the divided image of the divided image other than the encoding target divided image, a part or all of the image signal of the past or future frame of the divided image is added to enlarge the image size to be referred to, and the divided image of the enlarged current frame. And a means for performing coding after detecting a motion vector by using an enlarged divided image of a past or future frame, and an image signal coding device. 2. An image signal encoding apparatus for dividing an image signal of a current frame into some image data for a digital image signal and encoding each divided image signal,
Enlarge the reference image size by adding a part or all of the image signal of the past or future frame of the divided image other than the encoding target divided image to the image signal of the past or future frame of the encoding target divided image Means for detecting the motion vector for the divided image of the current frame by using the divided image of the past or future frame that has been enlarged, and encoding means for encoding the divided data of each divided image. And a unit for adding encoded data for enlarging the size. 3. An image signal of a current frame is divided into some image data with respect to a digital image signal, and coded by referring to an enlarged past or future frame divided image for each encoding target divided image. An image signal decoding device for decoding encoded data, which is a means for decoding the encoded data of each divided image, and a division other than the decoding target divided image into image signals of past or future frames of the decoding target divided image. A method of adding a part or all of the image signal of the past or future frame of the image to enlarge the image size to be referred to, and the divided image of the enlarged past or future frame from the decoded divided image information. A means for reproducing the divided image of the current frame including the enlarged area by using the means, and a means for deleting the enlarged area in the divided image from the reproduced divided image. An image signal decoding device characterized by.