JPH05236444A - Quantizing characteristic control method - Google Patents

Abstract

PURPOSE:To reduce the amount of transmission buffer memories and to frequently generate information concerning a part to use massive information by suppressing redundant information generation in respect to movement not to be visually followed up by enlarging a quantizing step size. CONSTITUTION:An input image signal is stored in a frame memory 1 and compared with the input image of the preceding frame as a block unit by a motion evaluation part 2. Then, difference between the position of a block having least error in the preceding frame and the position of the block of an input image signal is detected as a motion amount and transmitted to a motion compensation part 3, and a motion compensating block is read. At a subtraction part 4, difference between the block of the input image signal and the motion compensating block is calculated and transformed to the discrete cosine by a transformation part 5. The transformed signal is equally quantized by a quantizing step selected by a quantizing characteristic decision part 7 at a quantizing part 6 and transmitted through a transmission buffer memory 9 to a transmission line 10 after removing the redundancy of a quantizing signal sequence at a transmission coding part 8.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video coding technique for reducing the amount of communication information and transmitting it when performing video communication.

[0002]

2. Description of the Related Art In digital transmission of video signals, a motion-compensated interframe coding method has been known as a technique for highly compressing the amount of information to be transmitted. The motion-compensated interframe coding method uses video frame information before a video frame to be processed for information compression,
One of the interframe coding methods for predicting the video of the frame, detecting where the subject of the previous video frame is moving in the video frame, and based on the information, a highly accurate interframe prediction signal is obtained. Generating is the concept of the method.

In this method, the difference between the obtained inter-frame prediction signal and the input image signal is quantized. As for the applied quantization characteristic, the size of the quantization step is generally controlled according to the amount of use of the transmission buffer memory for sending to the transmission line at a constant speed.

[0004]

The present inventor has studied the above-described conventional motion compensation interframe coding method, and as a result, in this method, the visual acuity characteristic for the moving body of the visual sense is not utilized, so It has been found that the quantization step size may be reduced even for motions that cannot be followed, resulting in redundant information generation and increasing the amount of transmission buffer memory used.

The present invention provides an efficient quantization characteristic control method utilizing human visual characteristics in a motion-compensated interframe predictive coding method for detecting a motion of a subject and obtaining a highly accurate interframe prediction signal. To do.

The above and other objects and novel characteristics of the present invention will be apparent from the description of this specification and the accompanying drawings.

[0007]

In order to achieve the above object, the present invention divides an input image signal into blocks of a fixed number of pixels, and an image one or more fields before or one field or more after that block. From among the blocks of the same size with the highest similarity, the difference between the input image and the searched image is transformed using the orthogonal transformation matrix, and each signal after transformation is quantized and encoded. In the encoding method, the quantization characteristic to be used is determined by the motion amount between the input image and the image used for prediction within the range specified by the transmission buffer memory usage amount.

[0008]

According to the above-mentioned means, the quantization characteristic to be used is determined by the motion amount between the input image and the image used for prediction, so that the quantization step size can be set to a motion that cannot be visually followed. growing. Therefore, it is possible to suppress the redundant information generation, it is possible to reduce the use amount of the transmission buffer memory, it is possible to allow more information generation in the portion that uses a large amount of other information, This enables high-quality image transmission.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The configuration of the present invention will be described below together with an embodiment applied to a motion compensation interframe coding method for performing discrete cosine transform.

FIG. 1 shows a coding block section of a motion compensation interframe coding method for performing discrete cosine transform, to which the quantization characteristic control method of the present invention is applied. In the figure, 1 is a frame memory, 2 is a motion evaluation unit, 3 is a motion compensation unit, 4
Is a subtraction unit, 5 is a conversion unit, 6 is a quantization unit, 7 is a quantization characteristic determination unit, 8 is a transmission coding unit, 9 is a transmission buffer memory,
Reference numeral 10 is a transmission line, 11 is an inverse quantization unit, 12 is an inverse transformation unit, 1
3 is an adder.

The input image signal is stored in the frame memory 1 and, at the same time, is used for comparison with the input image signal of the previous frame in the motion evaluator 2 in units of blocks of 16 × 16 pixels. As a result of the comparison, the difference between the position of the block with the smallest error in the input image signal of the previous frame and the position of the block in the input image signal is detected as the amount of motion and sent to the motion compensator 3. Based on the result, the motion compensation block 3 reads the corresponding motion compensation block from the encoded / decoded video signal of the previous frame. Then, the subtracting unit 4 obtains the difference between the block in the input image signal and the motion compensation block, and the result is transformed by the discrete cosine transform in the transforming unit 5.

The signal converted by the conversion unit 5 is uniformly quantized in the quantization unit 6 by the quantization step selected by the quantization characteristic determination unit 7, and then the transmission encoding unit 8 is quantized.
Then, the redundancy of the quantized signal sequence is removed, and the quantized signal sequence is sent to the transmission line 10 via the transmission buffer memory 9. In addition, the signal quantized by the quantizing unit 6 passes through the inverse quantizing unit 11 and the inverse transforming unit 12, and in the adding unit 13, the motion compensating unit 3
Is added to the motion compensation unit 3 after being added.

FIG. 2 shows a quantization characteristic control method of the present invention in the quantization characteristic determination unit 7. In the figure, the horizontal axis represents the transmission buffer memory usage amount indicating the usage amount of the transmission buffer memory 9, and the point 27 represents the maximum buffer memory amount. The vertical axis represents the step size.
Two characteristic curves 23 and 24 are shown. In both cases, when the amount of transmission buffer memory used becomes large, the amount of transmission information generated is reduced by increasing the quantization step size, and conversely When the memory usage becomes small, the quantization step size becomes small and the transmission information generation quantity tends to increase.

In this example, when the quantization step size is changed according to the amount of transmission buffer memory used, the quantization step size is also changed according to the motion amount 25. The motion amount 25 indicates the positional deviation between the block to be coded and the block having the smallest error in the previous frame image, and is also transmitted to the receiving and decoding apparatus.

When the amount of movement 25 is large, the subject shows a large movement, and the human visual characteristics for a moving object deteriorate. Therefore, even if the quantization step size is increased, it is perceived as deterioration of the image. The degree of being treated is low. When the amount of movement 25 is small, the quantization step size is reduced to present sufficient quality for the agile visual characteristics. Therefore, at a certain transmission buffer memory usage amount point 26, the step size is to be changed within the illustrated range according to the magnitude of the motion amount.

However, if the quantization step size is determined only by the amount of motion, overflow and underflow of the transmission buffer memory 9 cannot be avoided, so the maximum and minimum values of the quantization step size are determined by the amount of transmission buffer memory used. , The quantization step size is determined by the motion amount 25 within that range. Characteristic curves 23 and 24 are
The range of the maximum value and the minimum value of this quantization step size is shown.

FIG. 3 shows an example of the configuration of the quantization characteristic determining unit 7 for implementing the quantization characteristic control method of FIG. 14,
15 is an input terminal, 16 is a quantization step size storage unit,
Reference numeral 17 is a minimum step size register, 18 is a range register, 19 is a limiter, 20 is an output terminal, 21 is a multiplication unit, and 21 is an addition unit.

The input terminal 14 has a transmission buffer memory 9
The amount of transmission buffer memory used is input to the quantization step size storage unit 16. A signal representing the amount of motion is input from the motion evaluation unit 2 to the input terminal 15 and input to the limiter 19. The quantization step size storage unit 16 is composed of a non-volatile memory, and the quantization step size at which the transmission buffer memory does not overflow or underflow with respect to each transmission buffer memory usage amount,
It is preset in the form of a minimum step size and an acceptable range.

The minimum quantization step size value is read from the quantization step size storage unit 16 in accordance with the usage amount of the transmission buffer memory, and is written in the minimum step size register 17. In addition, the allowable range value corresponding to the usage amount of the transmission buffer memory is the range register 1
Written in 8. When the amount of movement exceeds the upper limit value in the limiter 19, it is replaced with this upper limit value and output as a ratio value corresponding to the ratio with the limiter upper limit value. This output and the contents of the range register 18 are multiplied by the multiplication unit 21, and the result is added by the value of the minimum step size register 17 and the addition unit 22 to obtain the quantization step size, and the output terminal 20 Is output to the quantization unit 6.

With the above configuration, as is clear from FIG. 2, the quantization step size becomes large for a motion that cannot be visually followed. Therefore, since the redundant information generation is suppressed, the usage amount of the transmission buffer memory is reduced, and it is possible to allow more information generation to a portion that uses a large amount of information, so that high quality image transmission is possible. It will be possible.

Although the invention made by the present inventor has been specifically described based on the embodiments, the invention is not limited to the above-mentioned embodiments and various modifications can be made without departing from the scope of the invention. Needless to say. For example, the frame to be compared with the input image signal is the previous frame, but if the frame has already been encoded / decoded, the frame two or more previous or the frame one or more later is used. It is also possible. It is also possible to use the encoded and decoded video signals for motion evaluation in the motion evaluation unit 2.

Further, in the above description, the case where the invention made by the present inventor is mainly applied to the motion compensation interframe coding method for performing the discrete cosine transform, which is the field of use as the background, has been described. It is not limited.

[0023]

As described above, according to the present invention,
The amount of transmission buffer memory used is controlled by the quantization characteristic control method in the motion compensation interframe coding method by suppressing the redundant information generation by increasing the quantization step size for the motion that cannot be visually followed. It is possible to allow more information to be generated in a portion that uses a large amount of information, and it is possible to transmit a high quality image.

[Brief description of drawings]

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

FIG. 2 is a graph showing a quantization characteristic control method according to the present invention.

FIG. 3 is a block diagram showing a configuration of a quantization characteristic determination unit in FIG.

[Description of Codes] 1 ... Frame memory, 2 ... Motion evaluation unit, 3 ... Motion compensation unit, 4 ... Subtraction unit, 5 ... Transform unit, 6 ... Quantization unit, 7 ... Quantization characteristic determination unit, 8 ... Transmission code Quantization unit, 9 ... Transmission buffer memory, 10 ... Transmission path, 11 ... Inverse quantization unit, 12 ... Inverse conversion unit, 13 ... Addition unit, 14, 15 ... Input terminal, 16 ... Quantization step size storage unit, 17 ... Minimum step size register, 18 ... Range register, 19 ... Limiter,
20 ... Output terminal, 21 ... Multiplication section, 22 ... Addition section.

Claims (1)

[Claims] 1. An input image signal is divided into blocks of a fixed number of pixels, and a block of the same size having the highest degree of similarity is searched from an image one field or more before or one field or more after that block, and the input image signal is searched. In the motion compensation interframe coding method that transforms the difference between the searched image and the searched image using an orthogonal transformation matrix, and quantizes and encodes each transformed signal within the range specified by the transmission buffer memory usage amount. A method for controlling a quantization characteristic, characterized in that a quantization characteristic to be used is determined according to a motion amount between an input image and an image used for prediction.