JPH0545117B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an encoding method and apparatus for highly efficient compression encoding of moving image signals.

[Conventional technology]

Conventionally, IEEE Transactions on Communications (IEEE Transactions on Communications) has been used as a video signal encoding method that combines orthogonal transform encoding and interframe predictive encoding.
Communications) magazine, vol.COM−33, pp.1291
-1302, December 1985 (Reference 1) is known. The configuration of the encoding method described in Document 1 is shown in FIG. According to this encoding method, the encoding side consists of an orthogonal transformer 1, a subtracter 3, a quantizer 53, an adder 4, and a predictor 91, and the decoding side consists of an adder 11, a predictor 92, and an inverse orthogonal It consists of a converter 13.
In addition, 100, 300 are input terminals, 200, 40
0 is an output terminal. In this method, orthogonal transform coefficients are inter-frame encoded on the encoding side, and orthogonal transform coefficients are inter-frame decoded on the decoding side, and decoded images are obtained by performing inverse orthogonal transform on the orthogonal transform coefficients. When encoding inter-frame difference values of orthogonal transform coefficients, all orthogonal transform coefficients are quantized with the same quantization characteristic, and the amount of generated information is controlled by changing the step size of the quantization characteristic and the size of the dead zone. I'm going.

In addition, the 2nd International Technical Symposium on Optical and Electro-Optical Applied Science and Engineering (2nd
International Technical Symposium on
Optical and Electro Optical Applied Science
and Engineering）SPIE Conf.B594 Image
Coding, December 1985 (Reference 2) describes a coding method for orthogonal transformation coding of an interframe difference signal. The configuration of this encoding method is shown in FIG. According to this encoding method, the encoding side includes a subtracter 51, an orthogonal transformer 52, a quantizer 53, an inverse orthogonal transformer 54, an adder 55, a frame memory 56,
It consists of a predictor 57, a frame memory 58, a motion detection circuit 59, and a multiplexer 60, and the decoding side consists of a demultiplexer 61, an inverse orthogonal transformer 62, an adder 63, a frame memory 64, and a predictor 65. Note that 100, 300 are input terminals, 200,
400 is an output terminal. In this method, first, the orthogonal transform coefficients are divided into several sets and distributed.
By discarding orthogonal transform coefficients with small variances, deterioration in the image quality of the decoded image signal is suppressed as much as possible while encoding is performed at a constant bit rate. This orthogonal transform coefficient encoding method can be directly applied to the encoding of inter-frame difference values of orthogonal transform coefficients in Document 1.

[Problem that the invention seeks to solve]

In a coding method that performs interframe coding of orthogonal transform coefficients, by increasing the number of orthogonal transform coefficients to be discarded, encoding can be performed with a smaller amount of information. In conventional methods, it is determined whether to truncate orthogonal transform coefficients based on the magnitude of the interframe prediction error signal, but considering human visual characteristics, the resolution for moving parts may be low. By utilizing this characteristic, the amount of information can be further reduced. In other words, when the transform coefficients are truncated based on the magnitude of the interframe prediction error signal as in the prior art, all transform coefficients with significant interframe prediction error values are reduced to sign even in a motion region where the resolution may be low. must be transformed into The presence or absence of motion in the current frame can be approximately detected based on the magnitude of the interframe prediction error signal, but since the interframe prediction error signal includes noise due to encoding, the detected motion is noise. It does not necessarily correspond to the presence or absence of actual movement.

SUMMARY OF THE INVENTION An object of the present invention is to provide a more efficient method and apparatus for encoding a moving image signal by utilizing the human visual characteristic that the resolution of moving parts may be low.

[Means for solving problems]

The video signal encoding method of the present invention performs orthogonal transformation on an input image signal, and uses the obtained orthogonal transform coefficients and predictive orthogonal transform that is the orthogonal transform coefficient of the current frame predicted from the orthogonal transform coefficient of the previous frame. and the difference signal is used to locally decode the orthogonal transform coefficient of the current frame by calculating the sum of this difference signal and the predicted orthogonal transform coefficient to predict the orthogonal transform coefficient of the next frame. In a video signal encoding method that outputs a video signal as a coded image signal, it is possible to determine whether or not it is a moving region for each block of orthogonal transformation based on the inter-frame changes between the input image signal and the signal obtained by delaying the input image signal. The feature is that the orthogonal transform coefficients are truncated according to the determination result.

A first video signal encoding device of the present invention includes orthogonal transform means for orthogonally transforming an input image signal; orthogonal transform coefficients obtained by the orthogonal transform;
subtraction means for taking a difference from an output signal of an inter-frame prediction means to be described later; coefficient truncation means for cutting off some orthogonal transform coefficients of the output signal of the subtraction means in accordance with mode information provided by mode information formation means; addition means for calculating the sum of the output signal of the coefficient truncation means and the output signal of the interframe prediction means; interframe prediction means for predicting the next frame signal from the output signal of the addition means; and the coefficient truncation means. An output means for outputting an output signal as a coded image signal, and determining whether or not it is a moving area for each block of orthogonal transformation based on inter-frame changes between the input image signal and a signal obtained by delaying the input image signal. and mode information forming means for outputting mode information indicating which coefficient is to be truncated by the coefficient truncating means in accordance with the determination result.

A second video signal encoding device of the present invention includes orthogonal transform means for orthogonally transforming an input image signal, and a mode information forming means that provides some of the orthogonal transform coefficients obtained by the orthogonal transform means. coefficient truncation means for truncating according to mode information; subtraction means for taking the difference between an output signal of the coefficient truncation means and an output signal of interframe prediction means to be described later; and an output signal of the subtraction means and the interframe prediction means. addition means for calculating the sum of the output signal of the addition means; inter-frame prediction means for predicting the signal of the next frame from the output signal of the addition means; and output means for outputting the output signal of the subtraction means as an encoded image signal. , Based on inter-frame changes between the input image signal and the signal obtained by delaying the input image signal, it is determined whether or not it is a moving area for each block of orthogonal transformation, and depending on the determination result, which coefficient is selected by the coefficient truncation means. and mode information forming means for outputting mode information indicating whether to truncate.

A third video signal encoding device of the present invention includes orthogonal transform means for orthogonally transforming an input image signal; orthogonal transform coefficients obtained by the orthogonal transform;
subtracting means for calculating the difference between the output signal of the inter-frame prediction means to be described later; addition means for calculating the sum of the output signal of the subtraction means and the output signal of the inter-frame prediction means; inter-frame prediction means for predicting a signal of the next frame; coefficient truncation means for truncating some orthogonal coefficients of the output signal of the subtraction means according to mode information provided by the mode information forming means; An output means for outputting an output signal as a coded image signal, and determining whether or not it is a moving area for each block of orthogonal transformation based on inter-frame changes between the input image signal and a signal obtained by delaying the input image signal. and mode information forming means for outputting mode information indicating which coefficient is to be truncated by the coefficient truncating means in accordance with the determination result.

[Effect]

Whether each block of orthogonal transformation is a moving block that includes a moving part of the image or a stationary block that does not include a moving part of the image is determined by using the input image signal and a signal obtained by delaying the input image signal. This information is used to truncate more orthogonal transform coefficients for moving blocks than for stationary blocks. As a result, although the resolution of moving blocks decreases, it is possible to reduce the amount of information generated in moving blocks, and it is possible to encode moving image signals with a smaller amount of information and without significantly impairing subjective image quality. I can do it.

〔Example〕

FIGS. 1a, 1b and 1c are block diagrams respectively showing embodiments of the encoding device according to the present invention. The encoding apparatus of each of these embodiments is used to implement the encoding method of the present invention.

In each of the embodiments shown in FIGS. 1a, b, and c, the orthogonal transform coefficients are truncated at different locations.
The encoding device shown in FIG. 1a includes an orthogonal transformer 1 that orthogonally transforms an input image signal, and a subtracter 3 that takes the difference between the orthogonal transform coefficients obtained by the orthogonal transform and the output signal of the interframe predictor 2. , a coefficient truncation circuit 5 which truncates a part of the orthogonal transform coefficients of the output signal of the subtracter 3 according to the mode information given from the truncation coefficient determination circuit 81 , and a coefficient truncation circuit 5 which truncates the output signal of the coefficient truncation circuit 5 and the interframe predictor 2 . an adder 4 that calculates the sum with the output signal; an interframe predictor 2 that predicts the next frame signal from the output signal of the adder 4;
Based on the inter-frame changes between the input image signal and the signal obtained by delaying the input image signal, it is determined for each orthogonal transform block whether it is a moving area or not, and which orthogonal transform is selected in the coefficient truncation circuit 5 according to the determination result. The truncation coefficient determination circuit 81 outputs mode information indicating whether to truncate a coefficient.

In this encoding device, an image signal inputted from an input terminal 100 is orthogonally transformed in an orthogonal transformer 1, and then a difference between the image signal and the output signal of an interframe predictor 2 is calculated in a subtracter 3. The truncation coefficient determination circuit 81 generates mode information instructing whether to truncate the orthogonal transform coefficient. The mode information is input to the coefficient truncation circuit 5, and
It is output from the output terminal 210. A part of the orthogonal transform coefficients of the output signal of the subtracter 3 is truncated in the coefficient truncation circuit 5 based on the mode information output by the truncation coefficient determination circuit 81. The output signal of the coefficient truncation circuit 5 is outputted from an output terminal 200 as an encoded image signal. In the adder 4, the output signal of the coefficient truncation circuit 5 and the output signal of the interframe predictor 2 are summed. The interframe predictor 2 predicts the next frame signal based on the output signal of the adder 4 and outputs it to the subtracter 3 and the truncation coefficient determination circuit 81.

The encoding device shown in FIG. 1b includes an orthogonal transformer 1 that orthogonally transforms an input image signal, and a part of orthogonal transform coefficients obtained by the orthogonal transform that are truncated according to mode information provided by a coefficient determination circuit 81. A coefficient truncation circuit 8 that truncates, a subtractor 3 that takes the difference between the output signal of the coefficient truncation circuit 8 and the output signal of the interframe predictor 2, and a subtracter 3 that takes the difference between the output signal of the subtractor 3 and the output signal of the interframe predictor 2. An adder 4 that calculates the sum, an interframe predictor 2 that predicts the next frame signal from the output signal of the adder 4, and interframe changes between the input image signal and the signal obtained by delaying the input image signal. Accordingly, the truncation coefficient determination circuit 81 determines whether or not it is a moving region for each block of orthogonal transformation, and outputs mode information indicating which orthogonal transformation coefficient is to be truncated in the coefficient truncation circuit 8 according to the determination result. ing.

In this encoding device, an image signal input from an input terminal 100 is orthogonally transformed in an orthogonal transformer 1. The truncation coefficient determination circuit 81 generates mode information instructing whether to truncate the coefficient. The mode information is input to the coefficient truncation circuit 8 and output from the output terminal 210. In the coefficient truncation circuit 8, a part of the orthogonal transform coefficients is truncated based on the mode information output from the truncation coefficient determination circuit 81. The output signal of the coefficient truncation circuit 8 is subtracted from the output signal of the interframe predictor 2 by the subtracter 3, and then outputted from the output terminal 200 as an encoded image signal. In the adder 4, the output signal of the subtracter 3 and the output signal of the interframe predictor 2 are summed. The interframe predictor 2 predicts the next frame signal from the output signal of the adder 4 and outputs the predicted signal.

The encoding device of FIG. 1c includes an orthogonal transformer 1 that orthogonally transforms an input image signal, a subtracter 3 that takes the difference between the orthogonal transform coefficients obtained by the orthogonal transform and the output signal of the interframe predictor 2, An adder 4 that calculates the sum of the output signal of the subtracter 3 and the output signal of the interframe predictor 2, an interframe predictor 2 that predicts the next frame signal from the output signal of the adder 4, and a subtracter 3.
A coefficient truncation circuit 5 that truncates some orthogonal transform coefficients of the output signal according to the mode information given from the truncation coefficient determination circuit 81 and a frame-to-frame difference between the input image signal and the signal obtained by delaying the input image signal. It is comprised of a truncation coefficient determination circuit 81 that determines whether or not it is a moving region for each block of orthogonal transformation based on the change, and outputs mode information indicating which orthogonal transformation coefficient is to be truncated in the coefficient truncation circuit 5 according to the determination result. has been done.

In this encoding device, an image signal input from an input terminal 100 is orthogonally transformed in an orthogonal transformer 1, and then a difference between the image signal and the output signal of an interframe predictor 2 is calculated in a subtracter 3. Subtractor 3
The output signal of is summed with the output signal of the interframe predictor 2 in an adder 4. The interframe predictor 2 predicts and outputs the next frame signal based on the output signal of the adder 4. The truncation coefficient determination circuit 81 generates mode information instructing which orthogonal transform coefficient to truncate. The mode information is input to the coefficient truncation circuit 5 and output from the output terminal 210. In the coefficient truncation circuit 5, the truncation coefficient determination circuit 81
Based on the mode information output from the subtracter 3, some orthogonal transform coefficients of the output signal of the subtracter 3 are discarded. The output signal of the coefficient truncation circuit 5 is outputted from an output terminal 200 as an encoded image signal.

Two examples of the truncation coefficient determination circuit 81 are shown in FIGS. 2a and 2b.

The truncation coefficient determination circuit shown in FIG. The number of pixels equal to or greater than the threshold _Th1 is counted, and if the number of pixels is equal to or greater than the threshold _Th2 , the block is determined to be a moving block and is treated as another stationary block. Based on the mode information that is the result of this determination, the coefficient truncation circuit 5
For example, for a stationary block, see Figure 3a.
As shown in FIG. 3B, for the moving block, the coefficients marked with ◯ are preserved, and the orthogonal transformation coefficients marked with × are discarded.

In the truncation coefficient determination circuit shown in FIG. 2a, it is possible not only to determine whether a block is a moving block or a stationary block, but also to classify a moving block into several levels by preparing a plurality of threshold values _Th2 . In this case, the coefficient truncation circuit 5 prepares a plurality of mask patterns for moving blocks according to the classification, as in FIG. 3.

On the other hand, the truncation coefficient determination circuit shown in FIG. Prepare several signals, calculate the sum of absolute values of the differences between these signals and the signal of the current frame, and find the delay amount that minimizes the sum of absolute values. When the delay amount is at least one frame, the block is determined to be a stationary block; otherwise, it is determined to be a moving block. The amount of delay that deviates from one frame represents the amount of movement of the block from the previous frame, and it is also possible to classify moving blocks into several stages depending on the amount of movement.

Both of the above two truncation coefficient judgment circuits are
Although the result of determining whether a block is a moving block or a stationary block is used as mode information, it is also possible to use a truncated transform coefficient determination circuit in which a code indicating which orthogonal transform coefficient is to be truncated in coefficient truncation circuit 5 is used as mode information. can. Further, the mode information from the truncation coefficient determination circuit is transmitted to the terminal 21.
0 (see Figure 1), but the mode information does not need to be dedicated information that indicates which orthogonal transform coefficient to truncate, and may be a signal that can be derived from other encoded information. Therefore, in such a case, the terminal 210 is unnecessary.

By cutting off the transform coefficients in the coefficient truncation circuit 5 or 8 according to the characteristics shown in FIG. 3, the number of transform coefficients encoded in the moving block is reduced in the configurations of FIGS. Using visual characteristics, it is possible to reduce the amount of code while suppressing image quality deterioration. In the configuration shown in FIG. 1b, even if the block to be coded is a moving block, the high-frequency transform coefficients contained in the output signal of the interframe predictor 2 remain in the output of the arithmetic unit 3, so they must not be coded. However, among the transform coefficients output from the orthogonal transformer 1, high-frequency transform coefficients are discarded, so the amount of code can be reduced accordingly.

FIG. 4 shows an example of a decoding device corresponding to the encoding device of the present invention. This decoding device includes an adder 1
1, interframe predictor 15, coefficient truncation circuit 1
4. Consists of inverse orthogonal transformer 13, input terminal 300
The encoded image signal inputted from the interframe predictor 15 is summed with the output signal of the interframe predictor 15 in the adder 11 . In the interframe predictor 15, the adder 1
The next frame signal is predicted from the first output signal and output. The output signal of the adder 11 is subjected to inverse orthogonal transformation in the inverse orthogonal transformer 13 after a part of the orthogonal transform coefficients is discarded in the coefficient truncation circuit 14 based on the mode information inputted from the input terminal 310. Terminal 400 as a decoded image signal
It is output from

FIG. 5 shows a block diagram of an example of an encoding/decoding apparatus using the encoding apparatus of FIG. 1a and the decoding apparatus of FIG. 4. The encoding device has a quantizer 53, an inverse quantizer 71, and a multiplexer 60 added to the encoding device shown in FIG. 1a. The multiplexer 60 outputs the encoded image signal output from the quantizer 53, information on the quantization characteristics in the quantizer 53, and mode information output from the truncation coefficient determination circuit 81 to the output terminal 290.

Further, the decoding device has a demultiplexer 61 and an inverse quantizer 72 added to the decoding device shown in FIG. The demultiplexer 61 has input terminal 3
The signal from 90 is separated into an encoded image signal, quantization characteristic information, and mode information.

The encoding/decoding device in FIG. 5 uses the encoding device in FIG. 1a, but the encoding device in FIGS.
It is also possible to configure an encoding device by incorporating the multiplexer 60 and the multiplexer 60.

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

FIG. 6a shows that in the encoding device of FIG. 1a,
An example in which a motion compensation method is introduced as the interframe predictor 2 will be shown. The interframe predictor 2 includes 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 image movement from the input image signal and outputs it as motion information. The output signal of the adder 4 is sent to the inverse orthogonal transformer 31
is inversely orthogonally transformed. In the frame memory and variable delay circuit 33, the output signal of the inverse orthogonal transformer 31 is corrected by the amount of image movement detected by the motion detection circuit 34, and output as a signal for the next frame. Frame memory and variable delay circuit 3
The output signal of No. 3 is orthogonally transformed by an orthogonal transformer 32. The output signal of this orthogonal transformer 32 becomes the output image signal of the interframe predictor 2. Motion detection circuit 34 outputs motion information to multiplexer 60. The multiplexer 60 multiplexes the motion information with the encoded image signal and mode information and outputs the multiplexed motion information to the output terminal 290.

FIG. 6b shows an example in which a motion compensation technique is introduced as the interframe predictor 15 in the decoding apparatus of FIG. The interframe predictor 15 includes an inverse orthogonal transformer 41, an orthogonal transformer 42, a frame memory, and a variable delay circuit 43. A signal input from the input terminal 390 is separated into mode information and motion information as a coded image signal by the demultiplexer 61. The output signal of the adder 11 is inversely orthogonally transformed by an inversely orthogonal transformer 41 . The frame memory and variable delay circuit 43 corrects the output signal of the inverse orthogonal transformer 41 based on the motion information output from the demultiplexer 61, and outputs it as a predicted signal for the next frame. The output signal of the circuit 43 is orthogonally transformed by the orthogonal transformer 42 . The output signal of this orthogonal transformer 42 becomes the output signal of the interframe predictor 15.

〔Effect of the invention〕

In the encoding method and encoding apparatus of the present invention, high resolution is maintained for stationary blocks, and orthogonal transform coefficients are discarded for moving blocks to lower the resolution. The resolution of moving blocks is low due to human visual characteristics, so lowering the resolution of moving blocks will not result in a significant visual deterioration of the image quality.
By truncating orthogonal transform coefficients, a moving image signal can be encoded with a smaller amount of information.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the encoding device of the present invention, FIG. 2 is a block diagram showing an example of a truncation coefficient determination circuit, and FIG. 3 is an explanatory diagram of an example of the operation of the coefficient truncation circuit. FIG. 4 is a block diagram showing an example of a decoding device, FIGS. 5 and 6 are block diagrams showing an example of an encoding/decoding device using the encoding device of the present invention, and FIGS. FIG. 8 is an explanatory diagram of the conventional method. 1, 32... Orthogonal transformer, 2... Inter-frame predictor, 3... Subtractor, 4... Adder, 5, 8...
Coefficient truncation circuit, 81...Truncation coefficient determination circuit, 31...Inverse orthogonal transformer, 33...Frame memory and variable delay circuit, 34...Motion detection circuit, 60...Multiplexer.

Claims (1)

[Claims] 1. Orthogonal transform is performed on an input image signal, and the difference between the obtained orthogonal transform coefficient and the predicted orthogonal transform coefficient, which is the orthogonal transform coefficient of the current frame predicted from the orthogonal transform coefficient of the previous frame, is calculated. The sum of this difference signal and the predicted orthogonal transform coefficient is used to locally decode the orthogonal transform coefficient of the current frame to predict the orthogonal transform coefficient of the next frame, and the difference signal is used to predict the orthogonal transform coefficient of the next frame. In the encoding method of a moving image signal that is output as A method for encoding a moving image signal, characterized in that orthogonal transform coefficients are truncated according to the result. 2. orthogonal transformation means for orthogonally transforming an input image signal; orthogonal transformation coefficients obtained by the orthogonal transformation;
subtraction means for taking a difference from the output signal of the inter-frame prediction means; coefficient truncation means for cutting off some orthogonal transform coefficients of the output signal of the subtraction means in accordance with mode information provided by the mode information formation means; Adding means for calculating the sum of the output signal of the truncation means and the output signal of the inter-frame prediction means; Inter-frame prediction means for predicting the next frame signal from the output signal of the addition means; and the output signal of the coefficient truncation means. an output means for outputting the input image signal as a coded image signal, and determining whether or not it is a moving area for each block of orthogonal transformation based on inter-frame changes between the input image signal and a signal obtained by delaying the input image signal; An encoding apparatus for a moving picture signal, comprising mode information forming means for outputting mode information indicating which coefficients are to be discarded by the coefficient discarding means according to the result. 3 orthogonal transformation means for orthogonally transforming an input image signal; coefficient truncation means for truncating some coefficients of the orthogonal transformation coefficients obtained by the orthogonal transformation means according to mode information provided by the mode information forming means; and the coefficient truncation means. subtracting means for calculating the difference between the output signal of the means and the output signal of the inter-frame prediction means; addition means for calculating the sum of the output signal of the subtraction means and the output signal of the inter-frame prediction means; interframe prediction means for predicting the next frame signal from the output signal; output means for outputting the output signal of the subtraction means as a coded image signal; and an input image signal and a signal obtained by delaying the input image signal. mode information forming means for determining whether or not each block of orthogonal transformation is a moving region based on changes between frames; and outputting mode information indicating which coefficients to be discarded by the coefficient discarding means according to the determination result; An encoding device for a moving image signal, comprising: 4 orthogonal transformation means for orthogonally transforming an input image signal; orthogonal transformation coefficients obtained by the orthogonal transformation;
subtracting means for calculating the difference between the output signal of the inter-frame prediction means; addition means for calculating the sum of the output signal of the subtraction means and the output signal of the inter-frame prediction means; interframe prediction means for predicting a signal of a frame; coefficient truncation means for truncating a part of the orthogonal coefficients of the output signal of the subtraction means according to mode information provided by the mode information forming means; and an output signal of the coefficient truncation means. an output means for outputting the input image signal as a coded image signal, and determining whether or not it is a moving area for each block of orthogonal transformation based on inter-frame changes between the input image signal and a signal obtained by delaying the input image signal; An encoding apparatus for a moving picture signal, comprising mode information forming means for outputting mode information indicating which coefficients are to be discarded by the coefficient discarding means according to the result.