JPH06121301A - Motion compensation predictive coding device - Google Patents

Abstract

(57) [Abstract] [Purpose] Without using a large capacity buffer memory,
Further, it is possible to obtain a motion compensation predictive coding apparatus capable of intra-field coding without causing image quality deterioration. When the amount of data in the buffer memory 26 that holds the transmission rate of an output signal constant is small, the switch control circuit 30 turns off the switch 29 to cut off the prediction signal 20, and only the current frame image signal 4 is set to 1
By performing intra-field coding during the frame period,
Accumulation of errors in the transmission path is prevented without periodically turning off the switch 29 as in the conventional case.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a motion compensation coding apparatus which uses motion compensation based on a motion vector when coding a moving image.

[0002]

2. Description of the Related Art Conventionally, in the case of digitally transmitting image information, various encoding methods have been proposed in order to reduce the amount of transmission data.

One of the proposed encoding methods is
There is a method of encoding by switching between intraframe encoding and interframe encoding. Intra-frame compression is a method of reducing information by utilizing the characteristics of the same image in which adjacent pixels have similar brightness and color. In the actual image, most of the sky and walls have similar brightness and color, so
A compression of about 1/5 to 1/10 can be achieved by using only intraframe compression. The inter-frame compression is a method of obtaining an image by using only information for correction using similar images.

In a normal moving image, the patterns of adjacent frames are
They are similar with some movements and changes. Using this point, first, the similarity (motion, color, brightness, etc.) between the frame to be compression-encoded and the adjacent frame is calculated. Based on the calculation, the "predicted value", that is, the value of the frame closer to the "frame to be encoded" from the "proximity frame" is calculated.

Next, only the difference information from the "predicted value" is encoded (recorded / transmitted) from the frame to be encoded. Therefore, the amount of data (correction amount) is reduced. In other words, when a person moves to the right in a video in which only the person has moved, the predicted value is the pixel in which the person is in the previous frame, including the movement correction information, and the predicted value from all the pixels that have moved to the right. What subtracted is the difference.

It is known that, in the case of a conventional coding apparatus, generally, when an error occurs on a transmission line when the compression is performed by interframe processing, the error propagates. Therefore, when a predetermined number of interframe processings are performed, the intraframe processing is automatically performed.

[0007]

The amount of data per frame of each coded screen is 16K in the case of intra-frame processing.
It is about 25 Kbytes and about 7 to 10 Kbytes for inter-frame processing. That is, it is generally known that the intra-frame processing generates a larger amount of data than the inter-frame processing. Therefore, when performing intra-frame coding on a regular basis, if the quantization is performed using the quantization step immediately before that (quantization step used for inter-frame coding), the amount of data generated suddenly increases, and the transmission rate increases. It becomes an upper problem.

Therefore, if encoding is performed by changing the quantization step in order to suppress the amount of data generated, there arises a problem that the image quality is deteriorated this time. Also, DC
When the compression encoding using the T conversion is performed, the generated information is not biased for the DC (direct current) component of the intraframe compression process, and the information is entropy encoded (the probability of the generated information is high. Assign short codewords to things,
For those with a low probability, long codewords were allocated and the generated information could be reduced by encoding).

The above-mentioned problem occurs in a coding apparatus (at least having a predictive coding mode) which adaptively switches a plurality of coding modes to perform coding. In view of the background as described above, it is an object of the present invention to provide an encoding device that solves the above problems in the conventional encoding device and prevents signal deterioration.

[0010]

In the present invention, there is provided control means for executing intra-field encoding for one frame period according to the amount of data stored in the buffer memory.

[0011]

When the data storage amount in the buffer memory for keeping the transmission rate constant is large, the inter-frame coding is performed, and when the data storage amount is small, the intra-field coding is performed for one frame period. Intra-field coding is possible for one frame period without causing image quality deterioration.

[0012]

Embodiment An embodiment according to the present invention will be described below with reference to FIG.

In FIG. 1, 1 is an input terminal for inputting an analog image signal, 2 is an A / D conversion circuit for converting the input analog signal into a digital signal and blocking it, and 3 is a current frame of an encoding block. A prediction error calculation circuit that performs a difference operation between the image signal 4 and a prediction signal 20 described below to calculate a prediction error, 5 is an orthogonal transformation circuit that performs orthogonal transformation of the prediction error signal 6, and 7 quantizes the orthogonal transformation coefficient 8. It is a quantization circuit.

Reference numeral 9 is an inverse quantization circuit that inversely quantizes the quantized orthogonal transformation coefficient (prediction error signal) 10. Reference numeral 11 is an inverse orthogonal transformation circuit that inversely transforms the inversely quantized orthogonal transformation coefficient.
Reference numeral 12 is a reproduction image calculation circuit for calculating a reproduction image of the current frame by using the prediction error signal 13 and the prediction signal 20, reference numeral 14 is an image memory circuit for storing the reproduction image signal 15 of the current frame, and 16 is a reproduction image signal 17 of the previous frame. Is a motion compensation circuit for motion compensation, and 18 is a loop filter circuit for two-dimensionally low-pass filtering the motion compensation signal 19 and outputting a prediction signal 20.

Reference numeral 21 is a motion vector calculation circuit for comparing the current frame image signal 4 of the coded block and the reproduced image signal 17 of the previous frame to calculate the motion vector 22 of the coded block, and 29 is for motion compensation. A switch for forcibly setting the prediction signal 20 to "0", 30 is a buffer memory 26
The switch control circuit switches ON / OFF the switch 29 in accordance with the data storage amount to switch between inter-frame motion compensation predictive coding or intra-field coding for one frame period.

Reference numeral 23 is a variable length coding circuit for channel-coding the prediction error signal 10, 24 is a data synthesizing circuit for forming the transmission frame coded data 27 by the prediction error code 25 and the motion vector 22, and 26 is a transmission frame. A buffer memory for transmitting the encoded data 27 at a predetermined transmission speed, and 28 is an output terminal for outputting a transmission signal.

Reference numeral 31 is a quantization step control circuit which outputs a control signal 32 indicating a quantization step for controlling the data generation amount of the quantization circuit 7 and the inverse quantization circuit 9 according to the data amount of the buffer memory 26. .

Next, the operation of the above configuration will be described.

An analog image signal input to the input terminal 1 is converted into a digital signal by the A / D conversion circuit 2, and further, a block of M pixels in the horizontal direction and N lines in the vertical direction, for example M.
= N = 8 blocks. The divided current frame image signal 4 is input to the a input of the prediction error calculation circuit 3 and the motion vector calculation circuit 21. Prediction error calculation circuit 3
The predictive signal 20 is input from the loop filter 18 to the b input of the switch 29 when the switch 29 is ON, the difference operation with the current frame image signal 4 is performed, and the result is output as the predictive error signal 6.

The orthogonal transformation circuit 5 performs orthogonal transformation on the prediction error signal 6 from the prediction error calculation circuit 3 and outputs a prediction error orthogonal transformation coefficient 8. As the orthogonal transform method, in many cases, discrete cosine transform (DCT), which has a high conversion efficiency and is feasible for hardware implementation, is used.

The quantizing circuit 7 quantizes the prediction error orthogonal transformation coefficient 8 to calculate a prediction error orthogonal transformation quantized coefficient and outputs a prediction error signal 10.

At this time, the quantization circuit 7 changes the amount of data generated by changing the quantization step by the quantization step control circuit 31. This quantization step is controlled by the control signal 32 so as not to exceed the predetermined transmission rate output from the output terminal 28 by monitoring the amount of data stored in the buffer memory 26.

For example, when the amount of data stored in the buffer memory 26 is large, the quantizing circuit 7 is operated in a quantization step so that the amount of data generated is small, and the amount of data stored is reduced to a specified value. On the contrary, when the data storage amount is small, the quantization circuit 7 is operated in the quantization step so that the data generation amount increases, and the data storage amount is increased to the specified value.

On the other hand, the inverse quantization circuit 9 uses the prediction error signal 10
Dequantize. Since it is necessary to use the same quantization step as the quantization circuit 7 in the inverse quantization circuit 9 at this time, it is controlled by the quantization step control circuit 31 like the quantization circuit 7.

Next, in the inverse orthogonal transform circuit 11, the inverse quantized output is subjected to inverse orthogonal transform, and the prediction error signal 13 including the quantization / inverse quantization error is calculated. The reproduced image calculation circuit 12
The prediction error signal 13 including the quantization / dequantization error and the prediction signal 20 output from the loop filter 18 when the switch 29 is ON are added to calculate the current frame reproduced image signal 15 of the encoded block.

The image memory circuit 14 stores the reproduced image signal 15 of the current frame and outputs the reproduced image signal 17 of the previous frame.

The motion vector calculation circuit 21 compares the image signal 4 of the current frame with the reproduced image signal 17 of the previous frame stored in the image memory 14 to calculate the motion of the encoded block as a vector 22. .

The motion compensation circuit 16 motion-compensates the reproduced image signal 17 of the previous frame with the motion vector 22 and outputs it as a motion compensation signal 19. The loop filter circuit 18 performs two-dimensional low-pass filter processing on the motion-compensated coded block and outputs it as a prediction signal 20. The prediction signal 20 is input to the prediction error calculation circuit 3 and the reproduced image calculation circuit 12 when the switch 29 is ON.

On the other hand, the variable length coding circuit 23 has a prediction error signal (orthogonal transform quantization coefficient) 10 from the quantization circuit 10.
Of the prediction error code 25
Is output.

The data synthesis circuit 24 uses the motion vector 22.
Is encoded and combined with the prediction error code 25 to output transmission frame encoded data 27 in a predetermined format. The buffer memory 26 temporarily stores the transmission frame encoded data 27 and outputs it from the output terminal 28 at a predetermined transmission rate.

If an error occurs in the transmission path from the output of data from the buffer memory 26 to the input of the decoding device, the error is accumulated while operating as the motion compensation prediction interframe coding device. In order to proceed, it is necessary to turn off the switch 29 at an appropriate timing. Here, the switch control circuit 30 controls the switch 29 by monitoring the amount of data stored in the buffer memory 26.

For example, when the amount of data stored in the buffer memory 26 is large, the switch 29 remains ON and
It operates as a motion compensation predictive coding device between frames. When the amount of data stored in the buffer memory 26 is small, the switch 29 is turned off to perform intra-field coding for one frame period.

The switch 29 may be turned off once at a constant time in accordance with the error state of the data transmission path and in accordance with the amount of data stored in the buffer memory 26 as described above.
However, when the storage amount of the memory 26 is entirely exhausted, the switch 29 may be turned off within a predetermined time.

[0034]

As described above, according to the present invention, since the intra-field coding is executed according to the data storage amount of the buffer memory, it is not necessary to have a large buffer memory amount, and There is an effect that intra-field coding can be performed for one frame period without causing deterioration in image quality.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

[Explanation of symbols]

 26 buffer memory 29 switch 30 switch control circuit

Claims (1)

[Claims] 1. A motion compensation predictive coding apparatus for performing motion compensation between fields or between frames when coding an image signal, wherein interframe coding is performed in accordance with a data storage amount of a buffer memory for keeping a transmission rate constant. A motion-compensated predictive coding apparatus, characterized in that a control means for switching between the intra-field coding and the intra-field coding is provided.