KR20000010066A - Video data compression system using plural variable length coding tables - Google Patents

Abstract

PURPOSE: A video data compression system using plural variable length coding tables CONSTITUTION: The video data compression system includes a quadrature converting part (11) performing quadrature conversion of input video data; a quantizing part (12) quantizing the converted video data by using a certain scanning method and quantization table; a variable length coding part (13) producing video data as plural video data variable length coded; a control part (17) producing a control signal to select video data according to plural video data length. The high compressed data with short variable length code of plural variable length coded video data is automatically selected, thus making the high compressed video data processing possible.

Description

Image data compression device using multiple variable length encoding tables

The present invention relates to an image data compression apparatus using a plurality of variable length encoding tables. More particularly, the present invention relates to a plurality of variable length encoding tables that compress quantized image data by different variable length encoding tables and then compress the plurality of images. Variable length encoding processing that is more optimal for image data values in pixel block units input during image data compression by the variable length encoding process by controlling the data length to be selectively stored or transmitted by comparing data lengths. The present invention relates to an image data compression apparatus using a plurality of variable length encoding tables capable of automatically selecting, storing, or transmitting highly compressed image data.

In general, MPEG (Moving Picture Experts Group) (MPEG), a compression and decompression technique of image data, is performed through an encoding process, a compression process, and a decoding process, a compression process. This is done in three ways.

First, compression based on intra-screen (spatial) correlation, that is orthogonal transformation method using intra-spatial spatial frequency characteristics, and second, compression based on inter-screen (temporal) correlation, that is, motion vector obtained through comparison between front and rear screen information A motion compensation scheme used, and third, compression based on the bias of the data generation probability, that is, a variable length coding scheme in which a short code is assigned to a data value having a high probability of occurrence and a long code is assigned to a data value having a low probability of occurrence is used.

In the orthogonal transformation method, a discrete cosine transform (DCT) is most commonly used, and as shown in FIGS. 1 and 2, in the case of DCT, image data of one frame of a normal natural image (for example, The number of pixels = 352 X 288 (not shown) is divided into square blocks of 8X8 = 16 (bit) pixels, and the image data of the pixel block unit input is shown in FIG. The lower part has an inherent conversion characteristic of performing a one-sided conversion to video data of high frequency components.

As described above, the advantage of DCT is that pixel values (eg, luminance) that are generally irregularly distributed on the screen before conversion are unilaterally distributed as low-frequency component image data in the upper left part after DCT conversion. Therefore, the amount of video data can be compressed with little information loss by discarding the video data of unnecessary high frequency components distributed in the lower right part.

In order to discard the image data of the high frequency component, as shown in FIGS. 1 and 2, first, a scan operation is performed. In particular, a quantization processing operation is performed through a commonly used zigzag scan operation. In this case, the quantization is performed by dividing an arbitrary specific frequency value corresponding to each pixel of the image data by a preset quantization table, and dividing the division frequency value of the high frequency component by the division frequency value of the low frequency component. By setting a high value and rounding the remainder after division, a result value as shown in FIG. 2C can be obtained.

Therefore, the image data value after quantization is simply expressed as "C1 ~ C10 and the remaining 0".

On the other hand, the processing of the variable length coding scheme described above is as follows. As shown in FIG. 3, the quantized input image data values were, for example, 0,1,2,3, that is, 00,01,10,11, and the probability of generating each image data was 0.6, 0.2, 0.1 according to the experimental values. For example, assuming that the probability of occurrence is 0.1, 1 bit is allocated to 0 having the highest probability of occurrence, 2 bits are assigned to 1 having the highest probability of occurrence, and 3 bits are assigned to 2 and 3 having the lowest probability of occurrence.

Then, the average code length resulting from the variable length encoding process is computed as the average of the code length X allocation bits, so that the average code length is 1 X 0.6 + 2 X 0.2 + 3 X 0.1 + 3 X 0.1 = 1.6 (bits).

On the other hand, if the code length is fixed to 2 bits regardless of the probability of occurrence, 2 X 0.6 + 2 X 0.2 + 2 X 0.1 + 2 X 0.1 = 2 (bits).

Accordingly, it can be seen that the value (length) of the image data obtained by variable length encoding which gives variable allocation bits differently depending on the probability of occurrence is shorter, resulting in higher compression.

Meanwhile, the motion compensation method, which is a compression method based on the inter-screen (temporal) correlation described above, simplifies the image information by using a motion vector representing the degree of correlation between the screen information from the input front and rear screen information. The irrelevant bar will be omitted.

Such conventional techniques compress a large amount of image data more efficiently. In particular, the encoding process using the variable encoding method is currently used in most image compression apparatuses.

However, in order to set an optimal variable length coding table (VLC: hereinafter referred to as "VLC") for various natural images, an experimental value based on a large number of image data is required. The average user of a device using the technology has a problem that is virtually impossible.

Therefore, the present invention was created in order to solve the above problems, and includes a variable encoding table corresponding to various natural images, and includes a plurality of two or more variable encoding tables to perform a variable long coding process and to automatically generate shorter image data. It is an object of the present invention to provide an image data compression apparatus using a plurality of variable length encoding tables for controlling selection storage or transmission.

1 is a conventional image data compression apparatus by a single variable length encoding process,

2 illustrates a process of compressing and converting image data by a conventional image data compression apparatus.

Figure 3 shows a first embodiment of a conventional variable length encoding table

4 shows a first embodiment of an image data compression apparatus using a plurality of variable length encoding tables according to the present invention.

FIG. 5 illustrates another embodiment of compressing image data using a plurality of variable length encoding tables after separating and regrouping low frequency components and relative high frequency components after quantization conversion according to another second embodiment according to the present invention. One,

FIG. 6 conceptually illustrates a process of the regrouping operation of FIG. 5 according to another embodiment of the present invention.

※ Explanation of code for main part of drawing

1,11,101: Orthogonal Conversion Unit 2,12,102: Scanner

3,13,103: Quantizer (Q) 4,14,104: Quantizer

5,15 variable variable encoding unit 105,109 variable variable encoding unit

6,16,106: memory section 17,107: control section

108: Re-Ordering

An image data compression apparatus using a plurality of variable length encoding tables according to the present invention for achieving the above object is provided with a plurality of different variable encoding tables to convert the quantized image data into different variable length encoding tables It is characterized by controlling to automatically select, store, or transmit the length of the data, that is, the length of the code, that is, the highly compressed image data, among the plurality of image data that is encoded and output as a result.

In addition, the image data compression apparatus using the plurality of variable length encoding tables according to the present invention divides the low frequency component and the relatively high frequency component image data among the quantized image data values and regroups the image data in one pixel block unit. (Grouping, and further comprising a variable length coding table corresponding to the number of groupings and frequency components, and variable length coding processing as described above, and comparing the results to automatically select and store or transmit any one image data of short data length There is another feature in controlling as much as possible.

Hereinafter, a preferred embodiment of an image data compression apparatus using a plurality of variable length encoding tables according to the present invention will be described in detail.

FIG. 4 schematically illustrates a first embodiment of the present invention, wherein an orthogonal transform unit 11 orthogonally transforms image data input in pixel block units, and an arbitrary scan method and quantization table for the orthogonal transformed image data. A quantization unit 14 for quantizing using the quantization unit, a variable length encoding unit 15 for outputting the quantized image data as a plurality of variable length encoded image data by different variable length encoding tables, and And a control unit 17 for outputting a control signal to select any one image data according to a plurality of variable length encoded image data lengths.

An operation description of the image data compression apparatus using the plurality of variable length encoding tables according to the present invention configured as described above will be described in detail with reference to FIG. 4.

First, if any one pixel block of one frame image data (pixel: 352 X 288) of a natural image, that is, one word (Word, pixel: 8 X 8), is shown as an example, the spatial frequency characteristic of the image data is input. As shown in FIG. 2, low-frequency components and high-frequency components are biased into the upper left and lower right portions, respectively, by the orthogonal transform unit 11 performing the DCT operation.

Afterwards, the scanner 12 in the quantization unit 14 performs a zigzag scan operation and a quantization by a preset quantization table performed by the quantizer 13 to remove high frequency components of the image data distributed in different regions. The process converts and outputs the quantized image data having only the low frequency components as shown in FIG.

The image data having only the low frequency component is encoded by the variable length encoder 15 having at least two variable length encoding tables of FIG. 3 as described above.

In this case, the variable length encoder 15 has different variable length encoding tables VLC1 and VLC2 in FIG. 4 to simultaneously encode and process quantized transform image data which is output data of the quantization unit 14. As described above, the VLC1 and the VLC2 are set to have different numbers of bits of corresponding allocation codes on the premise that the occurrence probability of the input image data is different from each other, thereby outputting the same image data with variable length codes having different lengths. .

In this way, the output of the variable length encoder 15 for outputting two image data having different variable length code lengths is input to the memory unit 16 for storing and transmitting, wherein the memory unit 16 is stored. And controlling the transmission operation and comparing the average code lengths of the two video data subjected to the variable length encoding with each other to discriminate any one video data having a relatively short data length from the memory unit 16. It is controlled by the operation of the control unit 17 for controlling to be stored in.

5 and 6 schematically show a second embodiment as another embodiment of the present invention. An orthogonal transform unit 101 orthogonally transforms image data inputted to a pixel block. A quantization unit 104 to quantize using a scan method and a quantization table, a regrouping unit 108 to regroup into a plurality of groups according to frequency components of the quantized image data, and the plurality of regrouped images A first variable length encoding unit 105 for performing variable length encoding processing according to different variable length encoding tables according to frequency components and outputting the same as one image data, and converting the quantized image data into the first variable length. A second variable length encoding unit 109 for variable length encoding processing by another encoding unit and another variable length encoding table to output image data, and the first second variable. And a control unit 107 for outputting a control signal to select any one video data according to the plurality of video data lengths variable-coded by the long coding unit.

An operation description of the image data compression apparatus using the plurality of variable length encoding tables as the second embodiment of the present invention configured as described above will be described in detail with reference to FIGS. 5 and 6.

As shown in FIG. 5, the orthogonal transformed and quantized image data is input to the regrouping unit 108 shown as image data in which high frequency components have been removed and are unilaterally distributed only with low frequency components. Although the image data is low frequency components as a whole, there is relatively high frequency component image data between the image data.

For example, if the quantized image data is C1 to C10, C6 to C10 are relatively high frequency image data than C1 to C5. Thus, as shown in FIG. 6, the regrouping unit performs an operation of dividing the low frequency component and the high frequency component image data from the image data of one word and the image data of one word input thereafter.

In detail, two words having different frequency components are generated by dividing the image data corresponding to each word sequentially input into the image data of the low frequency component and the image data of the relative high frequency component.

The two words generated by regrouping are the first variable length encoder having VLC1 set based on the probability of occurrence of the image data of the low frequency component and VLC2 set based on the probability of occurrence of the image data of the high frequency component. Variable length encoding is performed by 105.

On the other hand, the quantized image data input to the regrouping unit 108 is subjected to variable length encoding by a second variable length encoding unit 109 having a conventional variable length encoding table, and each image obtained through such an operation. Data is input to the memory unit 106 for storage and transmission.

In this case, the memory unit 106 determines an average code length between the first image data which is the output data of the first variable length encoder 105 and the second image data which is the output data of the second variable length encoder 106. The control unit 107 controls the memory unit so as to discriminate image data having a relatively short code length and compare the image data stored in the memory unit to be image data having the short code length.

On the other hand, the delay time required for the grouping operation of the regrouping unit and the timing between the first and second variable-length encoders can be sufficiently implemented by a general tuning technique, and thus the detailed description thereof will be omitted.

An image data compression apparatus using a plurality of variable length encoding tables according to the present invention by the above-described configuration and operation is provided with at least two variable length encoding tables that are optimal for various natural images, and the result is variable length encoding. It is a very useful invention that greatly contributes to the improvement of the compression ratio of large-capacity image data by automatically storing and transmitting high-compression image data having a short average encoding length.

Claims (4)

An orthogonal transform unit which orthogonally transforms image data input in pixel block units; A quantization unit for quantizing the orthogonal transformed image data using an arbitrary scan method and a quantization table; A variable length encoder for outputting the quantized image data as a plurality of variable length coded image data by different variable length encoding tables; And a control unit for selectively outputting any one image data according to the plurality of image data sizes subjected to the variable length encoding process. The method of claim 1, And the controller selects and outputs one image data having a relatively small size among a plurality of variable length coded image data. An orthogonal conversion unit for orthogonally converting image data input in pixel block units; A quantization unit for quantizing the orthogonal transformed image data using an arbitrary scan method and a quantization table; A regrouping unit for regrouping into a plurality of groups according to frequency components of the quantized image data; A first variable length encoder for variable length encoding the plurality of regrouped image data using different variable length encoding tables and outputting the same as one image data; A second variable length encoder for variable length encoding the quantized image data by the first variable length encoder and another variable length encoding table to output image data; And a control unit for selectively outputting one arbitrary image data according to the plurality of image data sizes variable-length encoded by the first second variable length encoder. Data Compressor. The method of claim 3, The regrouping unit is divided into a low frequency component and a high frequency component and the image data compression apparatus using a plurality of variable length encoding table, characterized in that each re-grouping by one pixel block unit image data.