JPH08102940A - Moving picture processing unit and its method - Google Patents

Abstract

PURPOSE: To make deterioration in the image quality unremarkable by reducing each pattern size of a moving picture, adding a picture with high compression to its ridge so as to restore the entire size to an original image size before reduction thereby reducing the code quantity after compression. CONSTITUTION: An image whose size is reduced is given to an image composing section and a blank image is added to the surrounding of the image. The blank image data are image data given from a background image data part with high compression by an MPEG encoder 10. Thus, the increase in the code quantity after compression due to addition of the blank image is negligibly small. Thus, the moving picture data whose entire size is restored to the original size are given to the encoder 10, in which compression processing is implemented. Thus, the entire size of the moving picture entered to the encoder 10 is equal to a size of the original picture before reduction. Then a bit stream outputted from the encoder 10 is data corresponding to the original image size before reduction and a converter for system conversion processing in the case of display is not required.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus and method for compressing moving image data, and more particularly to an apparatus and method for reducing the amount of moving image data before compression.

[0002]

2. Description of the Related Art An apparatus for compressing a moving image by a predetermined method and outputting it at a fixed bit rate via a buffer has been provided or proposed. As such a device, for example, the predictive coding process between the screens is adaptively applied to the moving image data, the orthogonal transform process and the quantization process are further applied, and then the variable length encoding process is applied to the moving image code. There is an MPEG encoder that outputs as a bit stream conforming to the standardization MPEG. In the MPEG encoder, in order to prevent the overflow of the buffer, the target value of the code amount (target code amount) is set according to the free space of the buffer, and
By controlling the compression parameter (quantization step width or the like) according to the target code amount, the generated code amount is suppressed below the target code amount.

Japanese Unexamined Patent Publication No. 4-79587 discloses a moving image coding system for outputting coded data after compression at a variable rate, in which a future coded amount is set as a target coded amount based on a past averaged coded amount. A technique for controlling the quantization step width so as to converge is disclosed. In addition, in the technique of this publication,
When the quantization step width is increased as it is because the past code amount exceeds the allowable range, the image quality is extremely deteriorated.

In Japanese Patent Laid-Open No. 4-150279, when still image data is subjected to orthogonal transformation and variable-length coding to be compressed, the still image data before compression is thinned out to obtain a predetermined code amount after compression. The control technology is disclosed below. This publication also discloses a technique for controlling the code amount after compression to a predetermined amount or less by reducing the dynamic range of the peripheral portion of a still image and compressing the gradation range.

[0005]

In the control such that the generated code amount approaches the target code amount by varying the compression degree according to the free space of the buffer as in the MPEG encoder described above, there is a screen with a large data amount. If it continues for a long period of time, there is a problem that the degree of compression becomes very large and the image quality is significantly deteriorated, that is, the spatial resolution is reduced.

The above problem may be dealt with by, for example, dropping a top, but in this case, the time resolution is sacrificed at the expense of an unnatural shake (jerkiness) in the moving image. Occurs another problem. The present invention aims to provide another solution to the above problems.

[0007]

SUMMARY OF THE INVENTION The present invention is a moving image processing apparatus for compressing moving image data by a predetermined method and outputting it through a buffer at a fixed rate. A reduction means for reducing the number of pixels of each screen by reducing the size, and an image with a high degree of compression according to the predetermined method is added to the edge of each screen after the size reduction to reduce the size of each screen. And a synthesizing means for returning to the original size. The above-mentioned predetermined method is, for example, a compression method in which the predictive coding process is adaptively executed, the orthogonal transform process and the quantization process are further executed, and then the variable length coding process is executed. The method is used in the MPEG encoder.

In the above structure, means for monitoring the capacity of the buffer, means for activating the reducing means and the synthesizing means when the free capacity of the buffer becomes equal to or less than a predetermined threshold value,
May be added. More specifically, for example, a unit that sets a target value of the code amount after compression of each screen so that the buffer does not overflow according to the monitoring result, and controls each parameter of the compression process according to the target code amount. Means and
A means for deciding whether or not to operate the reducing means and the synthesizing means in accordance with the target code amount and switching the operation / non-operation may be added.

Further, according to the present invention, after reducing the size of each screen of a moving image, an image having a high degree of compression by a predetermined compression method is added to the edge of each screen to make the size of each screen the original size. This is a moving image processing method in which the image is returned and then compressed by the predetermined method. It is also a moving image processing method for outputting the compressed data at a fixed rate.

[0010]

Since the size of each screen of the moving image is reduced and the amount of data is reduced before data compression, the amount of data obtained by compressing each screen (reduced screen) of the moving image is not reduced. It is smaller than the amount of compressed data. Also,
After the size is reduced, an image is added to the edge and the image is returned to the original size (size before reduction), so that the size of the entire screen is not changed. Furthermore, since the additional image is an image in which the degree of compression in the compression method can be set to a very high level, such as a plain image having a low spatial frequency and small temporal variation, the additional image after being compressed by the additional image The amount of increase in the amount of data is slight.

[0011]

Embodiments of the present invention will be described below. FIG. 1 is a block diagram showing the configuration of the embodiment apparatus, FIG. 2 is a schematic diagram showing an original image, a reduced image and a composite image, FIG. 3 is a schematic diagram showing an example of the image reduction principle, and FIG. 4 is an MPEG encoder of FIG. 3 is a block diagram showing the configuration of FIG.

First, the MPEG encoder will be described with reference to FIG. The input moving image data is rearranged in the screen order in the preprocessing unit so that the screen referred to in the predictive coding process precedes, and is output to the switch SW11 in 8 × 8 pixel block units. In the switch SW11, the contact is switched to a or b according to the macro block type of the input block. That is, if it is an intra macroblock type for which only intraframe coding should be performed, a
The contact point is set and sent as it is to the DCT processing section.
Further, a forward prediction macroblock type in which a temporal interframe difference should be encoded, a backward prediction macroblock type in which a temporal reverse frame difference should be encoded, or a frame between a temporal forward direction and an average in the temporal reverse direction. B if it is a bidirectional prediction macroblock type for which the difference should be encoded
After being set at the contact point, the subtracter subtracts the difference from the image data in the frame memory, and then the difference is sent to the DCT processing unit.

The DCT processing section carries out DCT (discrete cosine transform) processing. The coefficient matrix after the DCT processing is input to the quantization unit and quantized. The step size of this quantization is controlled by the monitoring / encoding control unit. For example, when a screen with a large amount of data is continuously input and the remaining capacity of the buffer is reduced, and as a result, it is necessary to reduce the code amount input to the buffer, monitoring / encoding control is performed. The target code amount is switched to a smaller value and set by the unit. As a result, the quantization step width is increased, the degree of compression is increased, and the subsequent code amount is decreased. However, deterioration of image quality is inevitable.

The quantized data is then sent to the variable-length coding unit, variable-length coded (Huffman coded) using the bias of the appearance probability of the code, and then sent to the buffer. From the buffer, a fixed rate (eg 1.4 Mbp
s, 4 Mbps) bitstream of MPEG.

Further, in the above, if the picture to which the block belongs is an I picture or a P picture,
The coefficient matrix after quantization is a local decoding unit (inverse quantization unit,
The image data is also sent to the inverse DCT processing unit) and decoded, and then stored in either one of the frame memories (the frame memory in which the image data is already stored). In the case of the P picture, the image data of the frame memory in which the image data has been stored most recently is added by closing the switch SW14 and then stored in the frame memory. The image data stored in the frame memory in this way is sent to the above-mentioned subtractor (subtractor connected to the b contact of the switch SW11),
This is used for the predictive coding process described above.

Next, processing for reducing the screen size of a moving image will be described. In this device, the original image of FIG. 2 (a) is reduced as shown in FIG. 2 (b), and then a plain image is added to the peripheral portion (edge) as shown in FIG. 2 (c). Then, the image size of the whole (reduced original image + solid image in the peripheral portion) is returned to the original size before reduction and is input to the MPEG encoder.

The size reduction of the original image is performed by first inputting the original image to the horizontal filter section (see FIG. 1) and thinning out pixels in the horizontal direction. That is, (a) of FIG.
Pixels are thinned out as in (b). The image in which the pixels in the horizontal direction are thinned is next input to the vertical filter unit (see FIG. 1), and the pixels in the vertical direction are thinned out. This is shown in FIGS. 3 (b) and 3 (c). Although the pixel rows are thinned out as they are in FIG. 3, a known method such as synthesizing and replacing the thinned pixel rows with weighting can naturally be applied to the present invention.

The image thus reduced in size is then input to the image synthesizing unit (see FIG. 1) and a plain image is added to the peripheral portion. This plain image data is provided from the background image data section, and is the image data that is highly compressed by the MPEG encoder 10. Therefore, the amount of increase in the code amount after compression due to the addition of the plain image is negligible. In this way, the moving image data whose overall size has been restored to the original size is MPEG encoder.
The data is input to 10 and subjected to the above-described compression processing.

As described above, the overall size of the moving image input to the MPEG encoder 10 is equal to the size of the original image before reduction. Therefore, the bit stream output from the MPEG encoder 10 is also data corresponding to the size of the original image before reduction, and a converter for system conversion processing at the time of displaying on the display is unnecessary.

In the present embodiment, the MPEG encoder 10 determines whether to perform the size reduction processing and the plain image addition processing.
It is automatically switched based on the target code amount sent from the monitoring / encoding control unit (see FIG. 4) therein. That is,
As described above, in the MPEG encoder 10, the monitoring / encoding control unit monitors the capacity of the buffer, sets the target code amount so that the buffer does not overflow, and controls the quantization step width. This target code amount is also sent to the filtering process determination unit in FIG.

When the target code amount becomes smaller than a certain threshold value, the filter processing determination unit switches the contact of the switch SW1 to the contact b for a predetermined time to perform the size reduction processing and the plain image addition processing on the input moving image. Are controlled.
That is, as described above, when a screen having a large amount of data is continuously input, the MPEG encoder 10 sets the target code amount to a small value to increase the quantization step width, thereby reducing the subsequent code amount. However, in this case, if the quantization step width becomes too large, the deterioration in image quality becomes so large that it cannot be ignored. In the apparatus of this embodiment, the size reduction process and the plain image addition process are performed in such a case, so that the size of the original image is reduced, but the deterioration of the image quality is controlled to be inconspicuous. Since the target code amount is set for each of the I, P, and B pictures, for example, if any one of the target code amounts becomes smaller than the threshold value, the above switching can be performed. Good. It should be noted that the above switching may be controlled to be performed when all the target code amounts are smaller than the threshold value, or the control is performed to be performed when the target code amount of the I picture is smaller than the threshold value. You may. Further, the threshold value data is provided from the code amount threshold value data section.

In the above embodiment, the switch SW1 is switched according to the target code amount, but it may be controlled so as to be switched according to the quantization step width, or may be switched according to the remaining capacity of the buffer. You may control. Alternatively, the switching may be performed by an operator's operation input.

[0023]

As described above, according to the present invention, the size of each screen of a moving image is reduced before data compression to reduce the amount of data, and then an image with a high degree of compression is added to the edge of the entire image. Since the size of (reduced original image + edge) is returned to the original image size before reduction and the compression process is performed thereafter, the code amount after compression can be reduced. Further, after decoding, the size of the portion corresponding to the original image becomes smaller, but the deterioration of the image quality can be made inconspicuous.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an apparatus according to an embodiment.

FIG. 2 is a schematic diagram showing an original image, a reduced image, and a composite image.

FIG. 3 is a schematic diagram showing an example of an image reduction principle.

FIG. 4 is a block diagram showing the configuration of an MPEG encoder.

[Explanation of symbols]

 10 MPEG encoder

Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI Technical display area H04N 5/92 H04N 5/92 H

Claims (7)

[Claims] 1. A moving image processing apparatus for compressing moving image data according to a predetermined method and outputting the compressed moving image data via a buffer at a fixed rate by reducing the size of each screen of the moving image before compression. Reducing means for reducing the number of pixels of the screen, and synthesizing means for adding the highly compressed image in the predetermined method to the edge of each screen after the size reduction and returning the size of each screen to the original size. A moving image processing apparatus having. 2. The method according to claim 1, wherein the predetermined method adaptively performs predictive coding processing, then performs orthogonal transform processing and quantization processing, and further performs variable length coding processing. Is a video processing device. 3. The means according to claim 2, further comprising means for monitoring the capacity of the buffer, and means for activating the reducing means and the synthesizing means when the free capacity of the buffer is below a predetermined threshold value. And a moving image processing apparatus having. 4. The means for monitoring the capacity of the buffer according to claim 2, and setting a target value of the code amount after compression of each screen so that the buffer does not overflow depending on the monitoring result. And a means for controlling each parameter of the compression processing according to the target code amount, and a means for deciding whether or not to operate the reducing means and the synthesizing means according to the target code amount and switching the moving image. Image processing device. 5. After reducing the size of each screen of the moving image,
A moving image processing method, wherein an image having a high degree of compression according to a predetermined compression method is added to the edge of each screen to restore the size of each screen to the original size, and then compression is performed according to the predetermined method. 6. A moving image processing method for compressing moving image data according to a predetermined method and outputting at a fixed rate, wherein the size of each screen of the moving image before compression is reduced to reduce the number of pixels of each screen. And reducing the size of each screen to the original size by adding an image with a high degree of compression in the predetermined system to the edge of each screen after the reduction, and then performing compression in the predetermined system. Video processing method. 7. A moving image processing apparatus for encoding moving image data according to a predetermined method, wherein the size of each screen of a moving image before encoding is reduced to reduce the number of pixels of each screen. Means, a synthesizing means for adding a highly compressed image to the edge of each screen after size reduction at the time of encoding by the predetermined method, and returning the size of each screen to the original size, and returning to the original size. A moving image processing apparatus, comprising: an encoder that encodes each screen according to the predetermined method.