JPH0787583B2 - Video signal decoding device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to an apparatus for decoding a moving image signal which is compression-encoded with high efficiency.

[Conventional technology]

Conventionally, as a coding method of a moving image signal that combines orthogonal transform coding and inter-frame predictive coding, IEEE Transactions on Communications (IEEE Trans
actions on Communications), vol.COM−33, pp.1291
-1302, the one described in December 1985 (Reference 1) is known. FIG. 5 shows the configuration of the coding method described in Document 1. According to this encoding method, the encoding side has an orthogonal transformer 1,
It comprises a subtractor 3, a quantizer 53, an adder 4, and a predictor 91, and a decoding side comprises an adder 11, a predictor 92, and an inverse orthogonal transformer 13. In addition, 100 and 300 are input terminals, and 200 and 400 are output terminals. In this method, the coding side performs interframe coding of the orthogonal transformation coefficient, the decoding side performs interframe decoding of the orthogonal transformation coefficient, and performs inverse orthogonal transformation on this to obtain a decoded image. When encoding the inter-frame difference values of the orthogonal transform coefficients, all the orthogonal transform coefficients are quantized with the same quantization characteristic, and the amount of generated information is controlled by the step size of the quantization characteristic and the size of the dead zone. I am changing.

In addition, the 2nd International Technical Symposium on Optical and Electro-Optical Applied Science and Engineering
on Optical and Electro Optical Applied Science and
Engineering) SPIE Conf.B594 Image Coding, 1985 12
The month (reference 2) describes an encoding method for orthogonally transform-encoding an interframe difference signal. The structure of this encoding system is shown in FIG. According to this encoding method, the encoding side has a subtracter 51, an orthogonal transformer 52, a quantizer 53, and an inverse orthogonal transformer 5
4, adder 55, frame memory 56, predictor 57, frame memory 58, motion detection circuit 59, multiplexer 60, the decoding side demultiplexer 61, inverse orthogonal transformer 62, adder 63,
It is composed of a frame memory 64 and a predictor 65. In addition, 100 and 300 are input terminals, and 200 and 400 are output terminals. With this method,
First, the orthogonal transform coefficients are divided into several groups for distribution,
By rounding down the orthogonal transform coefficient with small variance,
While suppressing the deterioration of the image quality of the decoded image signal as much as possible,
It is devised so that encoding is performed at a constant bit rate. This encoding method of the orthogonal transform coefficient can be directly applied to the encoding of the inter-frame difference value of the orthogonal transform coefficient of Document 1.

FIG. 7 shows two examples of the coding device in which the coding method for rounding down the orthogonal transform coefficient shown in Document 2 is incorporated in the configuration of FIG. The coding apparatus of FIG. 7 (a) performs coefficient truncation inside the coding loop, and the coding apparatus of FIG. 7 (b) performs coefficient truncation outside the coding loop. These coding devices are composed of an orthogonal transformer 1, an inter-frame predictor 2, a subtractor 3, an adder 4, a coefficient truncation circuit 5, and a truncation coefficient determination circuit 81. The truncation coefficient determination circuit 81 is for determining the magnitude of energy of the inter-frame difference signal and determining which coefficient is to be truncated. In addition, 100 is an input terminal of an image signal, 200 is an output terminal of an encoded image signal, and 210 is an output terminal of mode information for instructing which coefficient is to be truncated.

[Problems to be solved by the invention]

The amount of information can be compressed by truncating the orthogonal transform coefficient to be encoded and reducing the number. At this time, in the conventional method, the orthogonal transform coefficient of the previous frame is decoded as the orthogonal transform coefficient of the current frame as it is with respect to the orthogonal transform coefficient coded with the inter-frame preliminary error being zero by truncation. Therefore, for example, the following problem occurs in the encoding / decoding device of FIG. 5 using the encoding device of FIG. 7 (a). FIG. 8 is a diagram for explaining the operation of the encoding / decoding device that uses one-dimensional orthogonal transformation. In the figure, the two waveforms in the square frame represent the waveforms corresponding to two of the orthogonal transform coefficients, and the waveform amplitude represents the magnitude of the coefficient.
The direct current component is omitted.

In FIG. 8, at the time of encoding the previous frame, a locally decoded signal such as waveform 508 is obtained, and orthogonal transform coefficients corresponding to waveform 506 and waveform 507, which are components of waveform 508, are sent to the decoding side and waveform 517 is generated. Is obtained as a decoded signal. In the case of using the method of taking the difference from the signal of the previous frame as a simple interframe coding, the predictor of FIG.
Waveform 506 as the prediction signal of the current frame output from 91
And the orthogonal transform coefficients corresponding to the waveform 507 are used. On the decoding side, the orthogonal transform coefficients corresponding to the waveform 515 and the waveform 516, which are the components of the waveform 517 as the output of the predictor 92 in FIG. Used. In the current frame,
When encoding an input signal such as the waveform 501, there is an inter-frame difference between the orthogonal transform coefficients corresponding to the waveform 502 and the waveform 503, which are components of the waveform 501, and the orthogonal transform coefficients corresponding to the waveform 506 and the waveform 507. Be taken. By this difference operation between frames, orthogonal transform coefficients corresponding to the waveform 504 and the waveform 505 are obtained. In the coefficient truncation circuit 5, the orthogonal transform coefficient corresponding to the waveform 504 is saved, and the orthogonal transform coefficient corresponding to the waveform 505 is truncated, so that the coefficients corresponding to the waveform 511 and the waveform 512 are encoded. In FIG. 8, the circle mark in the coefficient truncation circuit 5 indicates that the coefficient is stored, and the cross mark indicates that the coefficient is truncated. On the decoding side, the orthogonal transform coefficients corresponding to the waveform 515 and the waveform 516 predicted from the coefficient of the previous frame are output from the predictor 92 of FIG. 5, the orthogonal transform coefficient of the current frame is decoded, and the waveform 513 is obtained. An orthogonal transform coefficient corresponding to the waveform 514 is obtained. This results in waveform 513
And the orthogonal transform coefficient corresponding to the waveform 514 are inversely orthogonally transformed,
A decoded signal of waveform 523 is obtained. The orthogonal transform coefficients used for predicting the signal of the next frame are the orthogonal transform coefficients corresponding to the waveform 509 and the waveform 510 on the encoding side, and the orthogonal transform coefficients corresponding to the waveform 518 and the waveform 519 on the decoding side. Yes, and is input to the predictors 91 and 92 of FIG. 5, respectively. The decoded waveform 523 is converted into the input waveform 501 and the decoded waveform 51 of the previous frame.
Compared with 7, the decoded waveform 523 has a slope rising position shifted to the right from the input waveform 501 due to the effect of coefficient truncation, and has a shape close to the waveform 517 of the previous frame.

As described above, in the conventional method, the decoded waveform of the previous frame corresponding to the truncated coefficient remains in the decoded signal of the current frame as it is, resulting in a large deterioration in image quality. In particular, when the high-frequency coefficient of the frame difference is truncated, the high-frequency component of the object before moving, that is, the contour, remains in the background in a portion having a motion.

The object of the present invention is to eliminate such drawbacks of the conventional method,
An object of the present invention is to provide a decoding device for a moving image signal with less image quality deterioration by preventing unnecessary influence of the signal of the previous frame from remaining on the decoded image signal even if the orthogonal transform coefficient is truncated at the time of encoding.

[Means for solving problems]

A moving image signal decoding apparatus of the present invention includes an addition unit that adds an encoded image signal output by the encoding apparatus and an output signal of an inter-frame prediction unit described below, and an output signal of the addition unit that An inter-frame prediction means for predicting a frame signal, a coefficient truncation means for truncating a part of the orthogonal transform coefficients of the output signal of the addition means based on the mode information output from the coding device, and an output coefficient of the coefficient truncation means And an output means for outputting an output signal of the inverse orthogonal transform means as a decoded image signal.

[Action]

In FIG. 8, the output signal 523 of the decoding apparatus has a form close to the decoded signal 517 of the previous frame because the orthogonal transform coefficient corresponding to the waveform 516 which is the orthogonal transform coefficient of the previous frame is the current frame. The cause was that it was not corrected in interframe decoding.

Therefore, in the present invention, the orthogonal transform coefficient truncated on the encoding side is also subjected to interframe decoding in the decoding device,
It is truncated before the inverse orthogonal transform. This makes it possible to suppress deterioration of the decoded image signal.

〔Example〕

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

This decoding device includes an adder 11 for adding the encoded image signal output by the encoding device and the output signal of the interframe predictor 15.
And an inter-frame predictor 15 that predicts the signal of the next frame from the output signal of the adder 11, and truncates some orthogonal transform coefficients of the output signal of the adder 11 based on the mode information output by the encoding device. Coefficient truncation circuit 14 and inverse orthogonal transformer 13 for inverse orthogonal transforming the output coefficient of this coefficient truncation circuit
It consists of and.

In this decoding device, the encoded image signal input from the input terminal 300 is added to the interframe predictor 15 in the adder 11.
Is summed with the output signal of. The inter-frame predictor 15 predicts the signal of the next frame from the output signal of the adder 11. The output signal of the adder 11 is truncated in the coefficient truncation circuit 14 based on the mode information input from the input terminal 310, and is then subjected to inverse orthogonal transform in the inverse orthogonal transform circuit 13 and decoded. The converted image signal is output from the output terminal 400. As the mode information input from the terminal 310, in addition to information that directly indicates which coefficient is to be truncated, several patterns of coefficient to be truncated are prepared in advance, and the number of the pattern may be used as the mode information. it can. The mode information does not need to be dedicated information for instructing coefficient truncation, and may be a signal that can be derived from other encoded information.

FIG. 2 shows an operation explanatory diagram of a coding / decoding apparatus using the decoding apparatus of this embodiment and the coding apparatus of FIG. 7 (a). Similar to FIG. 8, this is an example in the case of using one-dimensional orthogonal transformation, and the two waveforms in the square frame show the waveforms corresponding to two coefficients of the orthogonal transformation coefficients. Indicates the magnitude of the coefficient. The direct current component is omitted.

2 will be described in comparison with FIG. 2 and 8
The difference in the figure is that in FIG. 8, the orthogonal transform coefficients corresponding to the waveforms 513 and 514, which are the output signals of the interframe decoding loop of the decoding device, are inverse-orthogonally transformed to output the output signal waveform 52.
However, in FIG. 2, after the orthogonal transform coefficient corresponding to the waveform 514 which is truncated on the encoder side in FIG. 2 is truncated by the coefficient truncation circuit 14, the orthogonal transform coefficient corresponding to the waveforms 520 and 521 is obtained. , Inverse orthogonal transform and output signal waveform
Got 522.

The waveform 522 is lower than the waveform 523, and is closer to the input waveform 501 at the position where the waveform crosses the level zero and crosses the level zero.
The difference in level from 01 is getting smaller. This tendency becomes more remarkable than the fact that the difference between the value of the coefficient truncated in the current frame and the value of the corresponding coefficient in the previous frame is large.

As described above, in the present invention, the deterioration of the decoded image signal can be suppressed as compared with the conventional method in which the coefficient is not truncated on the decoding side.

FIG. 3 shows a block diagram of an example of a coding / decoding device using the decoding device of FIG. 1 and the coding device of FIG. 7 (a). The decoding device has a demultiplexer 61 and an inverse quantizer 72 added to the decoding device of FIG. The demultiplexer 61 separates the signal from the input terminal 390 into an encoded image signal, quantization characteristic information, and mode information. Also,
The encoding device is the same as the encoding device of FIG.
Further, an inverse quantizer 71 and a multiplexer 60 are added. The multiplexer 60 is a coded image signal output from the quantizer 53, information on the quantization characteristic in the quantizer 53,
The mode information output from the truncation coefficient determination circuit 81 is output to the output terminal 290.

The encoding / decoding device of FIG. 3 is a case where the encoding device of FIG. 7 (a) is used, but similarly to the encoding device of FIG. 7 (b), the quantizer 53 and the dequantizer are used. The encoder 71 and the multiplexer 60 may be incorporated to form an encoding device.

An example of the inter-frame predictor 2 can be realized by simply delaying the input signal until the next frame and outputting. At this time, the inter-frame predictor 15 on the decoding side also simply delays the input signal until the next frame and outputs it.

In FIG. 4 (a), in the encoding device of FIG. 7 (a),
An example in which a motion compensation method is introduced as the inter-frame predictor 2 will be shown. The inter-frame predictor 2 is composed of an inverse orthogonal transformer 31, an orthogonal transformer 32, a frame memory and variable delay circuit 33, and a motion detection circuit 34. The motion detection circuit 34 calculates the amount of movement of the image from the input image signal and outputs it as motion information. The output signal of the adder 4 is inversely orthogonally transformed by the inverse orthogonal transformer 31. In the frame memory and variable delay circuit 33, the motion detection circuit is added to the output signal of the inverse orthogonal converter 31.
Only the amount of movement of the image detected in 34 is corrected and output as the signal of the next frame. The output signals of the frame memory and variable delay circuit 33 are orthogonally transformed by the orthogonal transformer 32. The output signal of the orthogonal transformer 32 becomes the output image signal of the interframe predictor 2. The motion detection circuit 34 outputs the motion information to the multiplexer 60. The multiplexer 60 multiplexes the motion information with the encoded image signal and the mode information and outputs the multiplexed information to the output terminal 290.

FIG. 4 (b) shows an example in which a motion compensation method is introduced as the interframe predictor 15 in the decoding apparatus of FIG. The inter-frame predictor 15 is composed of an inverse orthogonal transformer 41, an orthogonal transformer 42, a frame memory and a variable delay circuit 43. The signal input from the input terminal 390 is separated in the demultiplexer 61 into mode information and motion information as an encoded image signal. The output signal of the adder 11 is inversely orthogonally transformed by the inverse orthogonal transformer 41. In the frame memory and variable delay circuit 43, the output signal of the inverse orthogonal transformer 41 is corrected based on the motion information output from the demultiplexer 61, and the corrected signal is output as the prediction signal for the next frame. The output signal of the circuit 43 is orthogonally transformed by the orthogonal transformer 42. The output signal of the orthogonal transformer 42 is an interframe predictor.
15 output signals.

〔The invention's effect〕

In the conventional method, the signal waveform due to the conversion coefficient of the previous frame may remain in the decoded signal of the current frame as it is, which may cause a great visual deterioration. It can be suppressed and visual deterioration can be reduced.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a decoding device of the present invention, FIG. 2 is a diagram for explaining the operation of the embodiment of FIG. 1, and FIG. 3 is a diagram showing the decoding device of FIG. FIG. 4 is a block diagram showing an example of the encoding / decoding device used, FIG. 4 is a block section showing an example of an interframe predictor, FIGS. 5 and 6 are diagrams for explaining a conventional method, and FIG. FIG. 8 is a block diagram showing an example of the encoding device, and FIG. 8 is a diagram explaining the operation of the conventional system. 11 …… Adder 13,41 …… Inverse orthogonal transformer 14 …… Coefficient rounding circuit 15 …… Interframe predictor 42 …… Orthogonal transformer 43 …… Frame memory and variable delay circuit 61 …… Demultiplexer 72 …… Inverse quantizer

Claims (1)

[Claims] 1. A coded image signal output from a coding device,
Adder means for adding the output signal of the inter-frame prediction means described later, inter-frame predicting means for predicting the signal of the next frame from the output signal of the adder means, and based on the mode information output by the encoding device, Coefficient truncation means for truncating a part of the orthogonal transformation coefficients of the output signal of the addition means, inverse orthogonal transformation means for inverse orthogonal transformation of the output coefficients of the coefficient truncation means, and a decoded image signal for the output signal of the inverse orthogonal transformation means A decoding device for a moving image signal, comprising: