JPS6291091A - Adaptive type encoding system - Google Patents

Abstract

PURPOSE:To improve the coding efficiency by dividing a screen into small blocks, and selecting an inter-frame forecast encoding, an inter-field forecast encoding, or a within-field forecast encoding. CONSTITUTION:At a preceding field screen that is the output of a field memory 6, a preceding frame screen that is the output of a field memory 5, a block mean value calculator 7 and one line memory 8, the mean value of a block position just above a coding block that leads one block share of block line is obtained to be inputted to a dynamic vector detector 9 and a variable delay circuit 10. The dynamic vector detector 9 selects a block at the nearest of the present block out of the preceding field screen, the preceding frame screen, and the block positioned at just above, outputting it as a forecast block, and a difference from an input block is found at a subtractor 1 to quantize it by a quantizer 2. The dynamic vector detector 9 sends putting together which has been detected out of the preceding field screen, the preceding frame screen, and the mean value of the block positioned at just above and a moved part, as a piece of dynamic vector information to a receiving side with a forecasted error quantization value that is the output the quantizer 2.

Description

[Detailed Description of the Invention] [Summary] By dividing the screen into small blocks and selecting inter-frame predictive coding, inter-field predictive coding, or intra-field predictive coding according to the nature of the human-powered block signal, This improves encoding efficiency.

[Industrial application field]

The present invention relates to an improvement in an encoding method for image signals, particularly when a signal such as an image signal of a TV conference system is mostly in a still area and some parts move significantly.

In the above-mentioned encoding method, it is desired that encoding is performed that is suitable for the properties of the input block signal and that the encoding efficiency is high.

[Prior art and problems to be solved by the invention] Conventional encoding methods divide a screen into small blocks and perform predictive coding using the small blocks. There are two types: an inter-frame encoding method that uses a block at a certain position, and an intra-frame (inter-field) encoding method that uses the nearest block as the coded block one field before.
In the case of a still area, the coding efficiency is good because the same image is in the previous frame, but in the moving area, the coding efficiency decreases because the same image is not in the previous frame. Since the same thing does not exist in the screen one field before, the encoding efficiency in the still area is poor, and in either case, there is a problem that highly efficient encoding cannot be performed in both the moving area and the still area.

[Means for solving problems]

The above problem, as shown in the principle diagram of FIG.
The present invention uses, as a prediction block, a block that approximates the representative value of b) or the lining of an encoded block (c) that appears temporally before the encoded block (d) in the current field. The problem is solved by an adaptive coding method.

[Effect]

In the present invention, the block (a
), but if the motion is very large, intra-frame prediction may be better, and intra-frame predictive coding uses coded blocks surrounding the coded block. Focusing on the fact that the coding efficiency is superior when encoding a block, and in the case between the two, the encoding efficiency is superior when encoding between fields. The reference block in the previous field or the set of reference blocks in the previous field plane, or the block in the same field that is closest to the encoded block is used as the prediction block, so the prediction block is Encoding efficiency can be improved.

〔Example〕

FIG. 2 is a block diagram of an embodiment of the invention.

In the figure, 1 is a subtracter, 2 is a quantizer, 3 is an inverse quantizer, 4 is an adder, 5.6 is a field memory, 7 is a block average value calculator, 8 is a 1-block line memory, and 9 is an operating unit. A vector detector, 10 a variable delay circuit, and 11 a delay circuit, respectively.

Figure 2 shows an example where, in the case of intra-frame prediction, one block (c) directly above the current block is selected as the average value prediction for real-time processing and hardware simplification, as shown in Figure 1. be.

In the figure, the previous field screen which is the output of the field memory 6, the previous frame screen which is the output of the field memory 5, the block average value calculator 7 and the one line memory 8, are encoded one block line earlier. The average value of the block directly above the block is calculated using the motion vector information r
:, 9 and variable delay circuit 10, and the motion vector detector 9 detects the block closest to the current block from the previous field screen, the previous frame screen, and the block directly above. Either the field screen, the previous frame screen, or the average value of the block directly above is selected and output as a predicted block.

Then, the subtracter 1 calculates the difference from the large block, and the quantizer 2 quantizes and sends it out. The delay circuit 11 delays the input signal by a delay time until a prediction block is determined and output.

Since the motion vector detector 9 and the variable delay circuit 10 are described in Japanese Patent Application No. 59-44969, their explanation will be omitted.

At this time, the motion vector detector 9 detects the previous field screen,
Which of the average values of the previous frame screen and the block directly above was detected and the amount of movement are collectively transmitted as motion vector information to the receiving side together with the prediction error quantized value that is the output of the quantizer 2. .

Note that for the reference blocks within the field, the mode value may be used instead of the average value. Moreover, not only the block directly above, but also the block diagonally above, etc. may be used.

Next, we will discuss the effects of implementing the present invention in the third section.
This will be explained using FIGS. 4, 5, and 5.

FIG. 3 is a block diagram of the encoder, which differs from FIG. 2 in that a block validity/invalidity determination circuit 12 is added and a vector quantizer 13 is used as a quantizer.

Figure 4 shows the amplification of one person's face (large amount of movement), and Figure 5 shows the scene of a meeting with three people (medium amount of movement).
In order to perform low bit transmission, one field was encoded for every four fields.

In each figure, curve (1) is for interframe coding with motion compensation + intraframe average value prediction, curve (2) is for simple interframe + intraframe average value prediction, and curve (3) is for simple frame prediction. This is the case when there is a pause.

Also, the horizontal axis is entropy (bit/pel),
The vertical axis is S/N (db).

FIG. 6 is a diagram showing specifications.

In Figure 4, as the compression rate increases, the entropy decreases, but as can be seen by comparing curves (2) and (3), when intra-frame average value prediction is also performed, the S/N is large for the same entropy. Image quality improves.

Also, since the same S/N can be achieved with a smaller entropy, the compression ratio can be increased.

Furthermore, if motion compensation is also performed, the S/N ratio can be further improved and the entropy can be reduced.

The same effect can be obtained even when the amount of movement is medium as shown in FIG.

〔Effect of the invention〕

In this way, in general, interframe coding is used in the still area, when the motion is very large, the average value of the block directly above the block to be coded is used, and in the middle, either As a result, encoding efficiency can be improved from the moving domain to the static domain.

[Brief explanation of drawings]

Fig. 1 is a diagram showing the principle of the invention, Fig. 2 is a block diagram of an embodiment of the invention, Fig. 3 is a block diagram of a concrete encoder, and Figs. 4 and 5 are detailed explanations of the invention. Diagram 6 of the national policy for explanation is a diagram showing the specifications. In the figure, 1 is a subtracter, 2 is a quantizer, 3 is an inverse quantizer, 4 is an adder, 5.6 is a filled memory, 7 is a block average value calculator, 8 is a 1 block line memory, and 9 is a block line memory. a dynamic hector detector, and 10 a variable delay circuit, respectively. Notification of Japanese descent

Claims (1)

[Claims] When dividing a screen into small blocks and performing predictive coding on a small block basis, a block (a) in the previous frame or a reference block one field before the coding block (d) is used. set(
An adaptive encoding method characterized in that the block in (b) or the coded block (c) in the same field that is closest to the coded block is used as a prediction block.