JPS62206981A - Coding processing system - Google Patents

Abstract

PURPOSE:To code efficiently the identifying information of a valid/invalid block and to improve a transmitting efficiency by feeding one bit of an information quantity as valid/invalid block identifying information and executing the dividing processing only when the valid block is included. CONSTITUTION:The image signal inputted and digitized from a digital signal input terminal 1 is divided into the small blocks of the prescribed size in a scanning converting circuit 2. Further, a subtracting circuit 3 subtracts the forecasting value, namely, the output value of a frame memory 8. The inter- frame difference value is compared with the prescribed threshold in a valid block identifying circuit 4 and separated to a static area and a dynamic area. Only the signal of the dynamic area is scalar or vector-quantized for a block in a quantizing coder 5 and coded and transmitted from an output terminal 9.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an encoding processing method for highly efficiently encoding and transmitting inter-frame difference signals of moving images.

[Conventional technology]

Divide the inter-frame difference signal or motion-compensated inter-frame difference signal of a moving image into small blocks, identify whether or not the block should be transmitted according to the energy of the difference signal within the block, and then transmit the effective block directly or orthogonally. There are high-efficiency image coding methods that exchange scalar or vector quantization.

In these systems, information is compressed by giving one bit as valid/invalid block identification information and performing run-length encoding on the binarized data.

[Problem that the invention seeks to solve]

This is an attempt to improve the encoding efficiency of valid/invalid block identification information by taking advantage of the nature of moving images in which movement or brightness changes are local. Even if the number of valid blocks is concentrated to some extent and the number of valid blocks is small compared to the number of invalid blocks, if the valid blocks lack continuity, many short runs will appear even if run-length codes are used, so the effectiveness of information compression will be reduced. The drawback was that it wasn't enough.

In particular, if one tries to obtain a high information compression rate by changing the threshold value for identifying valid/invalid blocks, the above-mentioned drawback becomes more pronounced because valid blocks tend to appear scattered. There was a problem.

This invention was made to solve the above problems, and improves transmission efficiency by efficiently encoding identification information of valid/invalid blocks in a high-efficiency inter-frame differential encoding method for moving images. The purpose is to improve.

[Means for solving problems]

The first invention of the encoding processing method according to the present invention is to first transfer valid/invalid block information to several areas (
It divides into medium blocks) and allocates 1 bit to determine whether or not all of these divided areas are invalid blocks.Next, only if this medium block contains a valid block, further homomorphic or predetermined The area is divided into different shapes, and the amount of information is allocated in the same way. This operation is repeated until each block is finally scalar or vector quantized.

Further, the second invention according to the present invention is characterized in that the valid/SS diblock information of the inter-frame difference image before a predetermined frame time is predetermined for difference information from the valid/invalid block information of the inter-frame difference image at the current time. 1 bit is allocated depending on whether or not a valid block is included in these divided areas, and only when a valid block is included, the area is divided into several areas with the same shape or another predetermined shape. After dividing into regions, a similar amount of information is allocated, and this operation is repeated until each block is encoded by scalar or vector quantization.

[Effect]

1. According to this invention. The second invention is both effective/
One bit of information is given as invalid block identification information, and only when a valid block is included, the area is divided into the same shape or another predetermined shape, and the same amount of information is allocated.This operation continues, and scalar or Encoding is repeated up to the block unit to be vector quantized.

〔Example〕

FIG. 1 is a block diagram showing an embodiment of the present invention, in which 1 is a digital signal input terminal, 2 is a scan conversion circuit, 3 is a subtraction circuit, 4 is a valid/invalid block identification circuit, 5 is a quantization encoder, 6 is a quantization decoder, 7 is an adder circuit, 8 is a frame memory, and 9 is an output terminal.

Next, the operation will be explained.

A digitized image signal inputted from a digital signal input terminal 1 is divided into small blocks of a predetermined size in a scan conversion circuit 2.

Further, in the subtraction circuit 3, the predicted value, that is, the output value of the frame memory 8 is subtracted. The result of this subtraction, that is, the interframe difference value, is compared with a predetermined threshold value in the valid block identification circuit 4 and separated into a static area and a moving area.

Here, signals in the static domain are regarded as zero, and only signals in the dynamic domain are scalar or vector quantized for each block in the quantization encoder 5, and encoded and transmitted from the output terminal 9. Further, the inter-frame difference value reproduced by the quantization decoder 6 is added to the reproduced image of one frame time before in the adding circuit 7 and stored in the frame memory 8.

In the embodiments described above, it is necessary to encode and transmit valid/invalid block identification information in addition to the output information of the quantization encoder 5, and an example of encoding the valid/invalid block identification information will be shown below.

Figure 2 shows valid blocks as “1” and invalid blocks as “0”.
This is an example of valid/invalid block information when expressed as
Column (I) indicates valid/invalid block information, column (II) indicates code assignment, and column (m) indicates code string. (
If you use the method of repeating encoding by dividing into four squares (regions A, B, C, D) as shown in I), (1) Since region B does not contain any valid blocks, the code "Q II" is used. Abort allocation coding and use areas A, C, D
The code "1" is assigned to "1" and the process moves on to the next step.

(2) Divide area A into four parts: Aa, Ab, and Ac.

Ad. Areas Aa and Ab are all invalid blocks, and the code "0" is assigned to them and the encoding is discontinued. Area Ac,
The code "1" is assigned to Ad and the process moves to the next step. Area C,
Similar processing is performed for D as well.

(3) There are two valid blocks in area Ac and three valid blocks in Ad, so the codes of areas Ac and Ad are A, B, C, D.
If they are encoded according to the order assigned to them, each becomes “010
1”, '1110''. Similar processing is performed for small areas obtained by dividing areas C and D into four.

(4) The code of the valid/invalid block identification information becomes a code string as shown in FIG. 3 by defining the order of the areas as A, B, C, and D and arranging the codes in descending order of the area.

That is, (a) is information regarding areas A, B, C, and D;
(b), (c), and (d) are areas A and C.

Information regarding D, (e) and (f) are information regarding Ac and Ad territories, which is a further expansion of the information in (b), ()) and (ch) are information in area C, that is, (
The information in (c) is further developed, and (su) and (nu) are the information in the area, that is, the information in (d) is further developed.

Next, the second invention of this invention will be explained. In the second invention, invalid blocks are clustered to some extent (
This method takes advantage of the characteristics that appear in one frame as a collection. That is, a predetermined frame, for example 1 frame m
Difference information between the valid/invalid block information of the previous inter-frame difference image and the valid/invalid block information of the current inter-frame difference image (for example, exclusive OR if the valid/invalid block information is 1 bit) On the other hand, by applying the iterative region division method as in the first invention, the encoding efficiency can be further improved. This takes advantage of the fact that when the amount of motion in a moving image is small, there is a strong tendency for difference information to appear scattered. Note that if the number of divisions within a medium block is selected to be a power of 2, there will be no waste in allocating the amount of information.

〔Effect of the invention〕

As explained above, the first aspect of this invention. The second invention is
In a high-efficiency image encoding method, valid/invalid block information of the inter-frame difference image at the current time, or valid/invalid block information of the inter-frame difference image from a predetermined frame time ago, and validity/invalidity of the inter-frame difference image at the current time. Difference information from block information is divided into several regions, and codes are adaptively assigned to these divided regions, which improves encoding efficiency compared to conventional simple run-length encoding. There are advantages that can be improved.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a diagram showing an example of valid/invalid block information when a valid block is expressed as "1" and an invalid block is expressed as "0". FIG. 3 is a diagram showing an example of the obtained code string. In the figure, 1 is a digital signal input terminal, 2 is a scan conversion circuit, 3 is a subtraction circuit, 4 is a valid/invalid block identification circuit, and 5
is a quantization encoder, 6 is a quantization decoder, 7 is an addition circuit, 8
is a frame memory, and 9 is an output terminal. Figure 1 So, F 7 yoshiko

Claims (2)

[Claims] (1) After dividing the inter-frame difference signal or motion-compensated inter-frame difference signal of a moving image into small blocks, and identifying whether it is a valid block or an invalid block to be transmitted according to the energy of the difference signal within these blocks, In a high-efficiency image coding method that scalar or vector quantizes the difference signal included in the valid blocks block by block, valid/invalid block information for the entire frame or a part of the frame of the difference image is quantized in a predetermined number of formats. 1 bit is allocated depending on whether a valid block is included in these divided areas, and only when a valid block is included, the area is divided again into the same shape or another predetermined shape. An encoding processing method characterized in that a similar amount of information is allocated afterwards, and this operation is repeated for each block to be scalar or vector quantized. (2) After dividing the inter-frame difference signal or motion-compensated inter-frame difference signal of a moving image into small blocks, and identifying whether it is a valid block or an invalid block to be transmitted according to the energy of the difference signal within these blocks, In a high-efficiency image encoding method that scalar or vector quantizes the difference signal included in the valid block in units of blocks, the valid/invalid block information of the inter-frame difference image from a predetermined frame time ago and the inter-frame difference image at the current time are used. Difference information from valid/invalid block information is divided into several predetermined areas, and 1 bit is assigned depending on whether or not these divided areas contain a valid block. An encoding process characterized by dividing the region into isomorphic or predetermined different shapes, then allocating the same amount of information, and repeating this operation until the block is scalar or vector quantized. method.