JPH05304662A - Picture encoding and transmitting device - Google Patents

Abstract

(57) [Summary] [Structure] In a videophone and videoconference system encoding device, a motion vector obtained by motion compensation prediction is referred to, a large part of the motion vector is determined as a motion region, and further, this motion vector is determined. The area included in the area is defined as the moving area, and the rest is defined as the stationary area.
It is provided with a means for determining the threshold value in the valid / invalid discrimination. [Effect] The present invention has an effect of correctly encoding and reproducing a moving object with a high degree of attention.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image coding and transmitting apparatus for transmitting an image signal at a low bit rate, and more specifically, a band compression image coding for coding an image signal by motion compensation prediction and orthogonal transformation. And a transmission method.

[0002]

2. Description of the Related Art A conventional image coding and transmitting apparatus is disclosed in Japanese Patent Laid-Open No.
As shown in 51785, there is provided a means for determining a moving area and a still area based on the obtained motion vector and determining the threshold value for the valid / invalid judgment based on the result.

[0003]

However, in the above-mentioned conventional technique, the motion vector is easily detected and the edge of the moving object existing over a plurality of blocks is easily fixed in the moving region. In the central portion of, the motion vector is difficult to detect, and as a result, it is easy to be determined as a static area. For example, when a person is displayed, the outline of the face is determined as a moving area and is coded as an effective block, but the central part that makes the facial expression is determined as a static area and is not coded. Frequently appeared, and the facial expression was not reproduced correctly. Therefore, in order to solve such a problem, the present invention refers to a motion vector obtained by motion compensation prediction, determines a portion having a large motion vector as a moving region, and further, a region included in this moving region. Is defined as a moving area and the rest is defined as a stationary area, and the result is provided with a means for determining a threshold value for valid / invalid discrimination, so that a moving body with high attention can be correctly coded and reproduced. With the goal.

[0004]

According to the image coding and transmitting apparatus of the present invention, a means for sampling a video signal, the obtained data is divided into blocks, and the difference between the obtained block and the previous frame is predetermined threshold. An image code composed of means for discriminating between valid and invalid values, means for orthogonal transformation and quantization, means for motion compensation prediction with the previous frame, and means for transmission coding of information obtained by these. In the digitizing device, the motion vector obtained by the motion compensation prediction is referred to, a part having a large motion vector is determined as a moving region, a region included in this moving region is also determined as a moving region, and the rest is determined. It is confirmed as a static area, and this result is effective,
It is characterized in that it comprises means for determining a threshold value in the invalidity determination.

[0005]

EXAMPLES The present invention will be described in detail below with reference to examples.

FIG. 2 is a block diagram showing an outline of a general image compression coding apparatus. A video signal from the video camera 900, which is a signal source, is written in the frame memory 901.

Here, the intra-frame coding will be described first. The image data of the first frame read from the frame memory 901 for each block is judged by the block identifying means 913 as valid or invalid. The invalid block is not processed after the orthogonal transform. But,
In this case, since it is the first frame, all are valid blocks, and all are subject to the subsequent processing.

The in-frame mode data is orthogonally transformed by the orthogonal transformation means 902. Further, the obtained data is quantized by the quantizing means 903. The data transmitted through the communication path is rearranged by the data rearrangement means 904 by zigzag scanning, variable length coded one block at a time by the variable length coding means 906, and transferred to the communication protocol by the transmission coding means. , Sent to the transmission line.

The data quantized by the quantizing means 903 is prepared in preparation for the next interframe coding.
Inverse quantization is performed by the inverse quantization unit 907, and inverse orthogonal transformation is performed by the inverse orthogonal transformation unit 908. The obtained data is written in the frame memories 909 and 911 through the data path 931 in the in-frame mode.

Next, interframe coding will be described.

With respect to the image data of the first frame read from the motion compensation frame memory 911 in which the first screen is written, a motion vector is obtained by motion compensation with the frame memory A. This motion vector is encoded as transmission information, but at the same time, the image is divided into a moving area and a static area, and is also used as a block identifying means.

FIG. 1 is a block diagram of a circuit for discriminating between a moving area and a static area according to the present invention.

The frame memory B503 and the motion compensation frame memory 504 in FIG. 1 correspond to the frame memory B909 and the motion compensation frame memory 911 in FIG.

The image data stored in the frame memory B 503 and the motion compensation frame memory 504 is called for each block, and the motion vector detection circuit 505 detects the motion vector. The motion vector detected here is accumulated in the motion vector memory 506. When the processing for one frame is completed, the area discrimination circuit 50
2 and the area discrimination control circuit 501 determine the static area and the moving area.

FIG. 3 is a schematic diagram showing how the static area and the moving area are discriminated. First, a block 301 having a motion vector larger than a threshold value is determined as a moving area. This flag is stored in the area discrimination memory 507. When the determination for one frame is completed, the area discrimination control circuit 5
01, the area discrimination memory 507 is researched. Here, a portion 30 surrounded by a moving region in the horizontal direction
Add 2 to the moving area and confirm. That is, the block 301 determined as the moving area by the area determination by the motion compensation and the block 302 additionally determined as the moving area by the area determination result by the motion compensation are finally determined as the moving area. The remaining blocks are the blocks 303 determined to be the static area by the area determination by the motion compensation.

Returning to FIG. 2, the subsequent encoding will be described. The image data of the second frame read for each block from the frame memory 901 and the data output from the motion compensation frame memory 911 through the in-loop filter 910 are input to the subtraction device 920. At this time, the subtractor takes the difference between the first frame and the second frame.

Next, by the block identifying means 913,
It is judged as valid or invalid. The invalid block is not processed after the orthogonal transform.

The valid / invalid judgment is made based on the difference between the predetermined threshold value and the first frame and the second frame. In the present invention, separate threshold values are prepared for the static area and the moving area. .. The validated block is output to the orthogonal transformation unit 902, and thereafter, like the first frame, the orthogonal transformation unit 902, the quantization unit 903, the data rearrangement unit 904 by zigzag scanning, and the variable length coding unit 906 form one block. They are converted one by one, added to the communication protocol by the transmission encoding means, and sent to the transmission path.

The data quantized by the quantizing means 903 is further converted by the dequantizing means 907 and the inverse orthogonal transforming means 908 in preparation for the next interframe coding, and this time, in the frame memory 911. The contents are added together by the adder 921 and written in the frame memories 909 and 911.

Although this is repeated in the following frames, it may be advantageous to transmit in the intra-frame mode when the amount of code temporarily generated such as a scene change becomes significantly large. In such a case, the intraframe coding mode is forcibly inserted. However, in the intra-frame coding mode, the amount of generated code is generally large, and therefore the waiting time on the receiving side is long.

As described above, in the conventional example, all the generated codes are sequentially coded for each block and transmitted.

[0022]

According to the above configuration of the present invention, a motion vector obtained by motion compensation prediction is referred to, a portion having a large motion vector is determined as a moving region, and a region included in this moving region is determined. Is defined as a moving area, the rest is defined as a stationary area, and a means for determining a threshold value for valid / invalid discrimination is provided based on the result, so that a moving object with high attention is correctly coded and reproduced. Have an effect.

The above-described structure of the present invention can be widely applied to video telephones and video conference system coding devices.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an area discrimination circuit according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of an image coding and transmitting apparatus of the present invention.

FIG. 3 is a schematic diagram showing a state of area discrimination according to the present invention.

[Explanation of symbols]

301 block determined to be a moving area by area determination by motion compensation 302 block additionally determined to a moving area based on a result of area determination by motion compensation 303 block determined as a static area by area determination by motion compensation 501 area determination control circuit 502 Area discrimination circuit 503 Frame memory B 504 Motion compensation frame memory 505 Motion vector detection circuit 506 Motion vector memory 507 Area discrimination memory 900 Video camera 901 Frame memory A 902 Orthogonal transformation means 903 Quantization means 904 Data rearranging means by zigzag scanning 906 Variable length coding means 907 Inverse quantization means 908 Inverse orthogonal transformation means 909 Frame memory B 910 Loop filter 911 Motion compensation frame memory 912 Motion compensation data path 913 Block Data path click identification circuit 920 data path 931 frame mode subtractor 921 adder 930 frame mode

Claims (1)

[Claims] 1. A means for sampling a video signal, a means for dividing the obtained data into blocks, and a means for discriminating the difference between the obtained block and the previous frame by a predetermined threshold value. A motion vector obtained by motion-compensated prediction in an image coding device including a means for quantizing, a means for motion-compensated prediction with a previous frame, and means for transmission-encoding the information obtained by these means. , The area with a large motion vector is defined as a moving area, the area included in this moving area is also defined as a moving area, and the rest is defined as a stationary area. An image coding and transmitting apparatus comprising means for deciding a threshold in.