JPH06205396A - Image compression method with weighted screen - Google Patents

Abstract

(57) [Summary] [Object] The present invention divides a screen into a plurality of screens, and compares the block positions of the display screens to provide an image that provides a more vivid screen from the upper part of the screen to the central part than the lower part of the screen. A compression method, more specifically, an image compression method suitable for image transmission of a TV conference system or a TV telephone. [Structure] One screen is divided into a plurality of blocks by a screen division circuit, the block position where the divided block is displayed is detected by block position detection, and the quantization step width calculation circuit detects the quantum of each block by the block position. Calculate the conversion step width. It quantizes and encodes according to this width and sends it to the line via the transmission buffer circuit.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention divides a screen into a plurality of screens, compares the positions of blocks on the display screen, and intentionally quantizes the blocks at the bottom of the screen with coarse saturation, thereby changing the screen from the top to the center. More specifically, the present invention relates to an image compression method that provides a screen that is more vibrant than the lower part of the screen, and more specifically, to an image compression method that is suitable for image transmission in a TV conference system or a TV phone.

[0002]

2. Description of the Related Art As for image quality in a TV conference system or a TV phone, it is particularly important how to faithfully reproduce a human part rather than a peripheral part.

However, as in the conventional image compression method in the TV conference system, the motion vector for each block is calculated by comparing the difference between the data of the block obtained by dividing the screen into n × n and each block of the previous frame. Then, based on this, the inter-frame motion prediction method such as motion compensation and the optimal data amount for transmitting one frame from the line transmission rate and the number of transmission frames are assumed, and the quantization step size is based on this. In the method of calculating, quantizing, and encoding, the image is always compressed in one screen under common conditions.

FIG. 4 is a circuit block diagram showing an example of the conventional image compression method. An image taken in via the image pickup system is divided into n × n by the screen division circuit 1, stored in the frame memory 2 and sent to the motion prediction circuit 3.

The frame memory 2 is delayed by one frame and then input to the motion prediction circuit 3. Motion prediction circuit 3
Compares each block of the frame input from the screen division circuit 1 with the block at the same position in the previous frame from the frame memory 2 and the blocks around it, thereby detecting the change in the position between the frames and determining the motion vector. And

The inter-frame prediction circuit 4 includes a motion prediction circuit 3
From the value of the motion vector obtained from the above, the position of the block before the movement in the previous frame is detected, the difference data with that value is detected, and the image data is created by combining with the value of the motion vector.

The orthogonal transform circuit 5 performs a discrete cosine (DCT) transform on the difference data for each block.

The quantization circuit 6 quantizes the data from the orthogonal transformation circuit 5 with the step width indicated by the quantization step width calculation circuit 10.

The encoding circuit 7 variable-length-encodes the quantized value and the motion vector value, temporarily stores them in the transmission buffer 8, and then sends them out to the line at a constant speed according to the transmission rate. It

Here, the amount of data stored in the transmission buffer 8 is detected by the buffer storage amount detection circuit 9 and sent to the quantization step width calculation circuit 10.

[0011]

This image compression method is
The step size is calculated using the same function for each block in the screen, and a considerable amount of data is required for the image data of the unimportant part of the screen, resulting in an important image. There was a drawback that the amount of data representing a person was insufficient.

Generally, in a video conference system, the movement of the other person's eyes and mouth and the reproducibility of facial expressions are often particularly noted. How faithfully this part is reproduced has a great influence on the image quality evaluation result. Will be given. On the other hand, there is an upper limit to the amount of data that can be transmitted as one frame screen depending on the transmission system bit rate.

Therefore, the inventor of the present invention wondered whether it would be possible to perform high-definition image compression of only the central human portion in one image without changing the limited amount of transmission data.

In such a case, it is necessary to select a portion that requires high-definition image compression.

In the prior art, there is a method of recognizing the image data of the human part as a pattern, extracting only the human part from the input image information, and performing various processes.

However, this has a drawback that the pattern recognition is complicated, the processing circuit becomes large in scale, and enormous development cost and time are required.

The object of the present invention is to solve the above-mentioned drawbacks. Generally, in a TV conference, the face of a person is located above the central part of the screen, and at the bottom of the screen there is a conference desk, etc. It is an object of the present invention to provide an image compression method suitable for a TV conference system or a TV telephone system, which enables transmission of the blocks grouped in the central part from the lower part of the screen with higher definition.

[0018]

The present invention is intended to solve these problems, and one screen is divided into a plurality of blocks, which are calculated according to the transmission buffer accumulated amount vs. quantization step width characteristic. In the image compression method of quantizing based on the quantization step width, encoding and storing in the transmission buffer, a plurality of the transmission buffer storage amount vs. quantization step width characteristics are set to configure the grouped block group. By determining the block position to be set, the quantization step width is relatively small for blocks grouped from the upper part of the screen to the center part, and relatively large for blocks grouped at the bottom of the screen. It comprises a selection means for selecting a transmission buffer storage amount vs. quantization step width characteristic that gives a quantization step width.

[0019]

According to the above construction, the blocks grouped from the upper part of the screen to the central part can be quantized and coded so as to have higher definition than the lower part of the screen and transmitted.

[0020]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in more detail below with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of a circuit to which an image compression method according to the present invention is applied.

In the figure, the circuit parts given the same reference numerals as those used in FIG. 4 perform the same functions. The function selection unit 11 is set with a plurality of transmission buffer storage amount-quantization step width functions, and for example, g (x), f as shown in FIG.
It has a function of (x).

The function g (x) is associated with a small amount of transmission buffer storage and a quantization step width, and is a case of high-definition quantization according to the present invention.

The function f (x) is largely associated with the amount of transmission buffer storage and the quantization step width, and is a case where quantization is performed with a coarse definition according to the conventional method.

FIG. 3 shows an example of an image in which a person is actually displayed.
In many cases, an image is displayed on the screen that seems to attract attention. In this example, the face portion of the person and the hand portion of the person are displayed.

Now, one screen has 14 × 20 (n × n ′) = 2.
It is divided into 80 blocks. When one screen is displayed, the display is started from the left end of the upper part of the screen where n = 1 and n ′ = 1, and the block number of n is incremented. Then, when n increases to 14, the display is started from the second line of the screen, which is the next line, and n = 1 and n ′ = 2.
Becomes In this way, the block numbers of n and n'are incremented and displayed up to the bottom of the screen.

According to the present invention, as shown by the function g (x), the block number is determined by using the block position detector 12, and 1≤n, which is a block number corresponding to almost the center of the screen to the upper part of the screen. A block group up to '<16 is associated with a small amount in the transmission buffer storage amount, and a block group up to the block number 16≤n'≤20 corresponding to the lower part of the screen is a function of the transmission buffer storage amount. g
The quantization step width is associated with a relatively large value based on (x), and the quantization step width is associated with a large value based on the function f (x). As a result, data is sent from the transmission buffer circuit 8 to the line with high definition in the part corresponding to the upper part of the screen from the center part in one screen, and with coarse definition in the part corresponding to the lower part of the screen.

From the above, according to the present invention, the screen can be divided, and only the part corresponding to the upper part of the screen from the central part in one screen can be quantized and coded with relatively high image quality and transmitted.
It can contribute to the improvement of image quality in the V conference system and the TV phone.

[0028]

As described above, according to the present invention, in the image compression method in which the block position in one screen is detected, only the part corresponding to the upper part of the screen from the central part in one screen which is important for image evaluation is quantized. It is possible to perform high-quality transmission suitable for a TV conference system or a TV telephone by selecting a small step size and a large part corresponding to the lower part of the less important screen in one screen. There is an effect that it can be realized only by adding a relatively simple processing circuit.

[Brief description of drawings]

FIG. 1 is a circuit block diagram showing an embodiment of an image compression system according to the present invention.

FIG. 2 is a diagram showing an example of a quantization step calculation function selected by an image compression method according to the related art and the present invention.

FIG. 3 is a diagram showing an example of an image centering on a person, and is a diagram for explaining a quantization step width.

FIG. 4 is a circuit block diagram showing an example of a conventional image compression method.

[Explanation of symbols]

 1 Screen Division Circuit 2 Frame Memory 3 Motion Prediction Circuit 4 Interframe Prediction Circuit 5 Orthogonal Transformation Circuit 6 Quantization Circuit 7 Encoding Circuit 8 Transmission Buffer Circuit 9 Buffer Accumulation Detecting Circuit 10 Quantization Step Width Calculation Circuit 11 Function Selection Section 12 Block position detection

Claims (1)

[Claims] 1. An image in which one screen is divided into a plurality of blocks, quantized based on a quantization step width calculated according to a transmission buffer storage amount-quantization step width characteristic, coded and stored in a transmission buffer. In the compression method, a plurality of blocks grouped from the upper part of the screen to the central part are determined by setting a plurality of transmission buffer storage amount-quantization step width characteristics and determining the block positions that form the group of blocks. A selection means for selecting a transmission buffer storage amount vs. quantization step width characteristic that has a small quantization step width for a group and a large quantization step width for a block group grouped at the bottom of the screen is provided. An image compression method characterized in that the screen is weighted.