JPH05130587A - Buffer memory control circuit for image encoding device - Google Patents

Abstract

PURPOSE:To prevent the absolute storage quantity of a buffer memory from being exceeded by determining the step size of quantization according to the storage quantity of the buffer memory and the signal outputted from a cumulative addition part and outputting a control signal. CONSTITUTION:A frame memory 1 is stored with one frame of input digital image data P. A motion detection part 2 detects the motion of the image data P in block units according to one-frame delayed image data FD outputted from the memory 1 and the image data P. The signal outputted from the detection part 2 is accumulated and added by the cumulative addition part 3. A quantization control part 4 determines the quantization step size on condition that the storage quantity of the memory 1 is invariably less than a constant value, and outputs the control signal to a quantization part 12. The output of the quantization part 12 is encoded by an encoding part 13 and the encoded data are stored in the buffer memory 13 and outputted at a constant speed with the control signal of the control part 4. Consequently, even when data increase in amount abruptly, the absolute storage quantity of the memory 14 is prevented from being exceeded.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a buffer memory control circuit of an image coding device.

[0002]

2. Description of the Related Art An example of a buffer memory control circuit of a conventional image coding apparatus is shown in FIG. The buffer memory control circuit includes an orthogonal transform unit 11 that performs an orthogonal transform on the input digital image data P in units of blocks, and a quantizer 12 that quantizes the data output from the orthogonal transform 11. An encoding unit 13 that performs variable-length encoding on the data quantized by the quantizing unit 12 using a predetermined encoding table, and the encoded data output from the encoding unit 13 is accumulated to obtain a constant speed. The buffer memory 14 that outputs the data in the above-described manner, and the quantizing unit 12 based on the accumulated amount of the buffer memory 14 so that the accumulated amount is maintained at a certain amount.
And a quantization control unit 15 for determining the quantization step size of the above and outputting its control signal.

Next, the operation will be described. The input digital image data P is input in advance in units of blocks. Here, the block unit is M × N pixels.

First, in the orthogonal transformation section 11, the input data P
The orthogonal transformation is performed for each block of M × N pixels.
Next, in the quantization unit 12, a control signal for determining the quantization step size output from the quantization control unit 15 for the output data of the orthogonal transformation unit 11 (this will be described later).
Quantization is performed based on. Next, in the encoding unit 13,
Variable-length coding is performed on the output data of the quantization unit 12 according to a predetermined coding table. Then, the buffer memory 14 accumulates the encoded data output from the encoding unit 13 and outputs it at a certain speed.

In the quantization controller 15, the buffer memory 1
The quantizing unit 12 keeps the accumulated amount of 4 always at a certain fixed amount.
Set the quantization step size of. That is, when the storage amount in the buffer memory 14 increases, a control signal for coarsening the quantization step size is output to output the encoding unit 1.
3 decreases the amount of data output, and conversely, when the amount of storage in the buffer memory 14 decreases, a control signal for reducing the quantization step size is output and output from the encoding unit 13. It is designed to increase the amount of data.

[0006]

However, in the above-mentioned buffer memory control circuit of the conventional image coding apparatus, the quantization step size of the quantizer is determined only on the basis of the amount accumulated in the buffer memory. Therefore, it was not possible to cope with the rapid increase in the amount of data,
There was often the inconvenience that the buffer memory exceeded the absolute storage amount.

[0007]

SUMMARY OF THE INVENTION It is an object of the present invention to provide a buffer memory control circuit of an image coding apparatus capable of improving the disadvantages of the prior art and preventing the absolute storage amount of the buffer memory from being exceeded in spite of a sudden increase in coded data. To provide.

[0008]

A buffer memory control circuit of an image coding apparatus according to the present invention outputs an orthogonal transformation unit for orthogonally transforming input digital image data in units of blocks, and an output from the orthogonal transformation unit. A quantization unit that quantizes data, a coding unit that performs variable length coding on the data quantized by this quantization unit using a predetermined coding table, and an output from this coding unit. A buffer memory that accumulates encoded data that is output at a constant speed;
A quantizing control unit that determines the quantizing step size of the quantizing unit so that the accumulating amount is maintained at a certain amount from the accumulating amount of the buffer memory and outputs a control signal thereof is provided. Further, a frame memory for accumulating input digital image data for one frame is provided at the input stage of the orthogonal transformation unit, and one frame output from the frame memory is delayed at the input stage of the quantization control unit. A motion detection unit that detects the motion of image data in units of blocks from the image data and the input image data, and a cumulative addition unit that cumulatively adds the signals output from the motion detection unit in units of one frame are provided together. ing. Then, the quantization control unit has a function of roughly determining the quantization step size of the quantization unit in advance as needed based on the signal output from the cumulative addition unit. With such a structure, the above-mentioned object is to be achieved.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS. Here, the same reference numerals will be given to the same or equivalent components as those of the above-described conventional example, and the description thereof will be omitted.

FIG. 1 shows the configuration of an embodiment of the present invention. In the embodiment shown in FIG. 1, in the above-described conventional example, a frame memory 1 for accumulating the input digital image data P for one frame is provided side by side at the input stage of the orthogonal transformation unit 11, and the quantization control unit 15 is provided with the frame memory 1. In the input stage of the quantization control unit 4 provided in place of the above, motion detection of image data is performed in block units from the image data FD delayed by one frame output from the frame memory 1 and the input image data P. This is characterized in that a motion detecting section 2 for performing and a cumulative adding section 3 for cumulatively adding the signal DMIN output from the motion detecting section 2 in units of one frame are provided side by side. Then, the quantization controller 4 uses the buffer memory 1
The quantizing unit 12 keeps the accumulated amount of 4 always at a certain fixed amount.
In addition to the function of setting the quantization step size of, the quantization step size of the quantization unit 12 is roughly set in advance based on the signal output from the cumulative addition unit 3 as necessary. The configuration of the other parts is the same as the above-mentioned conventional example.

Next, the operation of this embodiment constructed as described above will be explained. Here, the input digital image data P is input in advance in units of blocks.

First, the input data P (here, M
It is assumed to be a block of × N pixels. ) Is stored in the frame memory 1 for one frame, and the frame memory 1 outputs the data FD delayed by one frame of the image. Next, the orthogonal transformation unit 11 performs orthogonal transformation on the data FD input from the frame memory 1 in block units of M × N pixels. Next, in the quantizer 12, the output data of the orthogonal transform unit 11 is quantized in accordance with the control signal of the quantum step size output from the quantization controller 4 (the details will be described later). To do. Next, the coding unit 13 performs variable length coding on the output data of the quantizing unit 12 according to a predetermined coding table. Then, the buffer memory 14 accumulates the encoded data output from the encoding unit 13 and outputs it at a certain speed.

In the present embodiment, the input data P and the output data FD of the frame memory 1 are also input to the motion detecting unit 2, and the motion detecting unit 2 blocks on the basis of these data. The movement of image data is detected in units.

The operation of the motion detector 2 will be described with reference to FIG. In FIG. 2, (1) is the input image data P, which is a block of M × N pixels. In FIG. 2, (2) is the image data FD one frame before the image data P, that is, the output of the frame memory 1. In the motion detecting section 2, based on the data P and the data FD, (3) in FIG.
The following calculation is performed for a certain range of the data FD as shown by.

[0015]

(Where i and j are -8≤i≤ + 7,-
8 ≦ j ≦ + 7. )

Therefore, the number of calculation results D for one block of the image data P is 16 × 16 = 256 according to the formula.
It becomes an individual. Since this calculation result D is the sum of the absolute values of the difference values in each pixel between the image data FD one frame before and the current image data P, i, at the minimum value of the 256 data, The value of j represents the position of movement of the current image data P with respect to the image data FD one frame before, and the motion detection unit 2 determines the minimum value DMI among them.
Output N.

The cumulative addition unit 3 cumulatively adds the output data DMIN of the motion detection unit 2 and constantly outputs the value. However, the value is reset to 0 by a frame clear pulse every one frame period.

The quantization control unit 4 determines the quantization step size so that the storage amount in the buffer memory 5 is always kept below a certain fixed amount, and sends the control signal to the quantization unit 12. That is, in the quantization control unit 4, as in the case of the quantization control unit 15 of the conventional example described above, the buffer memory 14 is always maintained so as to keep the accumulated amount of the buffer memory 14 at a constant amount.
When the amount of data stored in the buffer memory 14 increases, a control signal for coarsening the quantization step size is output to reduce the amount of data output from the encoding unit 13, and conversely, the amount of data stored in the buffer memory 14 decreases. If so, the encoding unit 1 outputs the control signal for reducing the quantization step size.
The amount of data output from 3 is increased. Further, in the quantization control unit 4, when the value of the output of the cumulative addition unit 3 exceeds a certain value, that is, when it is considered that there is a factor that the encoded data such as a scene change increases in the input image, the quantization is performed in advance. Set a coarse step size.

Therefore, in the present embodiment, when a factor such as a scene change of an image that causes an increase in coded data occurs, the quantization control unit 4 outputs a control signal that makes the quantization step size coarse in advance. Since the data is output to the encoding unit 12, the amount of data output from the encoding unit 13 decreases, and as a result, the storage amount of the buffer memory 14 decreases, so that the buffer memory 14 does not exceed the absolute storage amount. There is.

[0021]

As described above, according to the present invention, when a factor such as a scene change of an image that causes an increase in coded data occurs, the motion detecting section can detect this. Since the quantization control unit can coarsely set the quantization step size in advance by the output from the cumulative addition unit that cumulatively adds the output of the unit for each frame, the amount of data output from the encoding unit to the buffer memory is small. Therefore, it is possible to provide a buffer memory control circuit of an excellent image coding apparatus, which has not been available in the past, that can prevent the absolute storage amount of the buffer memory from being exceeded when the coded data rapidly increases.

[Brief description of drawings]

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

FIG. 2 is a diagram for explaining the operation of the motion detection unit in the embodiment of FIG.

FIG. 3 is a block diagram showing a configuration of a conventional example.

[Explanation of symbols]

 1 Frame Memory 2 Motion Detection Section 3 Cumulative Addition Section 4 Quantization Control Section 11 Orthogonal Transformation Section 12 Quantization Section 13 Encoding Section 14 Buffer Memory

Claims (2)

[Claims] 1. An orthogonal transform unit for orthogonally transforming input digital image data in units of a block, a quantizer for quantizing data output from the orthogonal transform unit, and a quantizer. An encoding unit that performs variable-length encoding on the quantized data using a predetermined encoding table,
A buffer memory that accumulates the encoded data output from the encoding unit and outputs the encoded data at a constant speed, and a quantum of the quantizing unit so that the accumulation amount is maintained at a certain amount from the accumulation amount of the buffer memory. In a buffer memory control circuit of an image encoding device, which includes a quantization control unit that determines a quantization step size and outputs a control signal thereof, in the input stage of the orthogonal transformation unit, the input digital image data is converted into one frame. A frame memory for accumulating data is provided side by side, and at the input stage of the quantization controller, image data of a certain block is output based on the image data delayed by one frame output from the frame memory and the input image data. A motion detection unit that performs motion detection and a cumulative addition unit that cumulatively adds the signals output from the motion detection unit on a frame-by-frame basis are provided side by side. An image coding apparatus characterized in that the quantization control unit has a function of coarsely setting a quantization step size of the quantization unit in advance based on a signal output from the cumulative addition unit, if necessary. Buffer memory control circuit. 2. The motion detection unit detects the pixel data of the image data one frame before and the current image data from the image data delayed by one frame output from the frame memory and the input image data. 2. The buffer memory control circuit of the image coding apparatus according to claim 1, having a function of calculating and outputting the minimum value of the sum of absolute values of the difference values.