JPH10136384A - High efficiency encoding system for moving image - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve image quality and to reduce a transmission information amount by adaptively using the encoding of a frame block and a field block. SOLUTION: Input television signals are the data of the frame block constituted of an odd-numbered field and an even-numbered field and decomposed into the frame block and the field block by a frame/field converter 12. Then, the frame block is inputted to a frame encoder 13 and the field block is inputted to a field encoder 14. Entropy E which is the output of the encoders 13 and 14 and a local decoding screen R are compared in a comparator 16 and a selector 15 selects a Huffman code ZH which is the encoded result of one of the frame encoder 13 and the field encoder 14 by the result.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an encoding system such as discrete cosine transform (DCT) and vector quantization of a moving image.

[0002]

2. Description of the Related Art Moving picture communication such as a video conference and CD-
In a high-efficiency video coding method for storing a video in a ROM or the like, one frame or field unit is used.
Each screen is divided into blocks of, for example, 16 pixels × 16 lines, and high-efficiency coding is performed using intra-frame coding or inter-plane coding for coding a difference between a reference screen and a current screen by motion compensation. Is going. In FIG. 4, reference numeral 41 denotes a subtractor, which obtains a difference between the input screen X1 and the prediction screen X2 to generate a prediction error screen X3. 42 is a discrete cosine transform (DC
T), 43 is a quantizer, 44 is an inverse quantizer, 45 is an inverse discrete cosine transform (IDCT), 46 is an adder and IDCT
The local decoding screen R is generated by adding the prediction error screen X5 and the prediction screen X2 restored by the above. Reference numeral 47 denotes a predictor, which uses intra prediction, motion compensation prediction, inter prediction, and the like. The output X4 from the quantizer 43 is encoded by the Huffman encoder 48, and outputs a Huffman code H and an encoded information amount (entropy) E.

[0003]

In the above-described coding apparatus, coding is performed in block units on a frame screen or a field screen. In this case, (1) When there is only a frame block, the correlation between the upper and lower lines is smaller for an image having motion in intra-frame coding and for an image having an acceleration motion in inter-frame coding. , The coding efficiency decreases. (2) In the case of only field blocks, the coding efficiency is low for an image with no motion in the intra-frame coding because the correlation between the upper and lower lines is lower than that in the frame unit. Since the vertical position is different from that of the above, the prediction accuracy is lower than the prediction in frame units, and the coding efficiency is lower. Accordingly, an object of the present invention is to improve the coding efficiency and the image quality, which are the drawbacks of the above-mentioned conventional method.

[0004]

In order to solve the above-mentioned problems, the present invention provides a first difference between a locally decoded image obtained by local decoding after frame encoding and an original image, and a local decoding after field encoding. Comparison result obtained by comparing the magnitude of the second difference between the locally decoded image and the original image obtained as a result, the first encoded information amount after frame encoding, and the second code after field encoding. Based on the second comparison result comparing the sizes of the encoded information amounts, adaptively select between frame encoding processing in units of frame blocks or field encoding processing in units of field blocks. A feature is that a flag signal indicating a block unit in the selected encoding process is added to a block in the selected encoding process and transmitted.

[0005]

According to the present invention, when encoding is performed on a pixel block basis, encoding such as DCT or vector quantization is performed on a frame block and a field block, and an encoded information amount (entropy) as an encoding result is obtained. In addition, a block is adaptively selected (selected) based on a difference (error) between the locally decoded image and the original input image, and a flag indicating the type of the selected block is added to each block and transmitted. .

Therefore, the present invention adaptively selects a frame block or a field block so as to increase the coding efficiency, and transmits the frame block or the field block together with a flag indicating which one has been selected. Therefore, the coding efficiency is improved and the object of the invention is achieved.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a first embodiment of the present invention will be described with reference to FIG. In the figure, an input television signal 11 is data of a frame block composed of an odd field and an even field, and is decomposed into a frame block and a field block by a frame / field converter 12. However, a field block is input to the field encoder 14. The entropy E output from the encoders 13 and 14 and the local decoding screen R are compared by a comparator 16, and as a result of the comparison, the selector 15 selects one of the frame encoder 13 and the field encoder 14. Huffman code Z
Select H.

FIGS. 2A and 2B show the structure of block data of a frame signal and a field signal, respectively. The size of the block to be processed is 1 for the luminance signal.
Six pixels × 16 lines and two color signals are each set to 8 pixels × 16 lines, and these are collectively called a macroblock. In the block of the frame signal, as shown in FIG.
Further, even-numbered lines have even-numbered field data (x). In the field signal block, as shown in FIG. 2B, the upper eight lines have odd field data (デ ー タ), and the lower eight lines have even field data (×). FIG. 4 (conventional example) is a specific configuration of the encoders 13 and 14 in FIG. In the encoding of each macroblock, the encoding process from the DCT unit 42 including the Huffman encoder 48 to the IDCT unit 45 is performed every 8 lines × 8 pixels, and the Huffman code H
And the encoded information amount E are output. The local decoding screen R is output from the adder 46. These outputs are output by the frame encoder 13 of FIG. 1 to HR, ER, RR,
And the field encoder 14 outputs H
Output as F, EF, RF.

At the time when the processing of each macroblock is completed, the comparator 16 compares the sum of the coded information amount of one macroblock and the local decoded value in the case of frame coding and field coding. For example, to compare the amount of coded information, the entropy of the luminance signal is compared, a coding method with a small amount of coded information is set as a candidate, and for the comparison of the local decoded value, the cumulative sum of the difference square value with the input pixel is calculated. Each of them is obtained and an encoding method with a small error (difference) is set as a candidate. As an example of the selection, when the error (difference) between the locally decoded image and the original image is small and the amount of coded information is small, the field coding method is set as a candidate, and otherwise, the frame coding method is set as a candidate. That is, the encoded information amount of the frame is ER, the encoded information amount of the field is EF, and the error between the locally decoded image and the original image in the frame encoding is DR.
(Not shown), when the error between the locally decoded image and the original image in the field coding is DF (not shown),
When F <ER and DF <DR, the field coding method is used, and the others are frame coding methods. Comparator 16
The Huffman code output is selected by the selector 15
And the selection flag is output as ZM. As the type of the prediction screen, an intra prediction, an inter prediction, and a motion compensation prediction screen can be considered. Therefore, as the encoding mode including the prediction method, there are six modes of intra-frame encoding, intra-field encoding, inter-frame encoding, inter-field encoding, motion compensation inter-frame encoding, and motion compensation inter-field encoding. It becomes possible.

In the decoder, a flag Z is set for each macroblock.
A decoding process is performed on a frame or a field block based on M.

Various modifications are possible in practicing the present invention. For example, in the above-described embodiment, DCT is used as a method of encoding an input image signal. However, the present invention is not limited to this, and various encoding methods such as vector quantization and DPCM can be applied to the present invention. Also, the size of the block is not limited to 16 pixels × 16 lines, but may be 32 pixels × 16 lines.
Various sizes, such as 32 lines, are adaptable.

The comparator 16 shown in FIG. 1 can make a comparison based on the result of either the local decoded value or the encoded information amount.

As a further modification, the encoder 1 of FIG.
Instead of encoding a frame signal and a field signal using two encoders as shown in 3 and 14, instead of performing a frame signal or a field signal selection in advance at the input stage of the encoder 34 as shown in FIG. There is a method of selectively performing either one of the encoding methods. Hereinafter, a case of a block of 16 pixels × 16 lines will be described.

First, in the case of intra-frame encoding, an input screen X1
Has 16 pixels × 16 by the frame / field converter 31
A line frame block RX and a 16-pixel × 8-line field block LX are created and input to the comparator 32. Here the standard deviation RS for the frame block
D and the sum LS of the standard deviations of the two field blocks
D is obtained (the standard deviation is obtained with respect to the absolute value of the difference between the pixel value and the average value in the block). These results are input to the selector 33, where the block having the smaller standard deviation is selected and output as the selection flag ZM.
The encoder 34 performs frame or field encoding based on the selection flag ZM.

In the case of inter-plane coding, an input screen X1
The frame / field converter 31 creates a frame block RX and a field block LX for each screen of the predicted screen X2 created by the encoder and outputs it to the comparator 32. The comparator 32 calculates the standard deviation from the absolute value of the difference signal between the input screen and the motion-compensated prediction screen (by local decoding) stored in the frame or field encoder 34 for the frame block and the field block, respectively. And LSD as selector 3
Output to 3. Hereinafter, selection and encoding are performed in the same manner as in the case of intra-frame encoding.

In this modified example, since the encoding process is performed only once, it is possible to reduce the size of the hardware. However, as described in the first embodiment, the decoding screen and the input screen are compared and the amount of encoded information is increased. Since the selection based on the comparison of the frame / field coding is not performed, the coding efficiency may be reduced as compared with the first embodiment.

Instead of the standard deviation, it is possible to use a deviation of a block such as a variance or a sum of absolute values of difference signals.

[0018]

As described above, in the present invention, the coding efficiency of the conventional frame signal only coding and the field signal only coding is reduced by adaptively using the coding of the frame block and the field block. To improve the image quality and reduce the amount of transmitted information. An example of the effect is the ISO test video (flower gar
For d-en, football), the image quality (S / N ratio) was improved by 0.2 to 0.3 dB and the information transmission amount was reduced by about 5% in the CCIR601 image format at a bit rate of 4 Mbit / s.

[Brief description of the drawings]

FIG. 1 is an apparatus configuration diagram to which an embodiment of the present invention is applied.

FIGS. 2A and 2B are illustrations for explaining a block which is a unit to be processed. FIG. 2A shows a configuration of a frame block, and FIG. 2B shows a configuration of a field block.

FIG. 3 shows an apparatus configuration diagram to which a modification of the present invention is applied.

FIG. 4 shows a detailed configuration diagram of a conventional encoder used in the present invention.

[Explanation of symbols]

 Reference Signs List 11 input screen 12 frame / field converter 13 frame encoder 14 field encoder 15,33 selector 16,32 comparator 31 frame / field converter 34 frame / field encoder 41 subtractor 42 DCT 43 quantizer 44 Inverse quantizer 45 IDCT 46 Adder 47 Predictor 48 Huffman encoder

Claims (2)

[Claims] 1. A moving picture is subjected to transform coding processing such as discrete cosine transform or vector quantization in units of blocks, and adaptively applied in units of frame blocks in macroblocks or in units of fields in its own field in macroblocks. In a high-efficiency coding method, a first difference between a locally decoded image obtained by local decoding after frame coding and an original image, a locally decoded image obtained by local decoding after field coding, and an original image And a second comparison result of comparing the magnitude of the first encoded information amount after frame encoding and the magnitude of the second encoded information amount after field encoding. Based on the comparison result of, the frame encoding process for each frame block unit,
Or adaptively selecting field encoding processing for each field block unit, and adding a flag signal indicating the block unit in the selected encoding process to the selected block in the encoding process and transmitting the block. A high-efficiency coding method for moving images characterized by the following. 2. The method according to claim 1, wherein the first comparison result is a first difference and a second difference is a second difference.
Is selected to perform encoding in units of field blocks only when the second comparison result is smaller than the difference between the first encoded information amount and the second encoded result is smaller than the second encoded information amount. Item 2. A high-efficiency encoding method for a moving image according to Item 1.