JP2803768B2 - Image processing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method for compressing input image data and decompressing the compressed image data by a high efficiency coding method using DCT (Discrete Cosine Transform). The present invention relates to an image processing method that requires a small memory capacity for processing, improves the efficiency of the processing, and suppresses image deterioration during decompression processing.

[0002]

2. Description of the Related Art In recent years, DCT has become LS
Communication, computers and AV
In such fields, the DCT has become a leading role of the high efficiency coding system.

Accordingly, various manufacturers have proposed various DCT-dedicated LSIs. In the image processing method using the DCT, for example, 8 × 8 pixels are defined as one block, and the DCT operation is performed in units of the one block,
Facilitates image processing for compressing input image data,
Further, an inverse DCT operation is performed on the compressed data to facilitate image processing for expanding the compressed data.

[0004]

However, in the above-described image processing, the data of the DCT operation result is larger than the input image data. For example, when the input data is 8 bits, the output data is 12 bits. However, there is a problem in that the memory capacity in the above-described compression processing becomes large, and the image processing is troublesome and time-consuming, that is, the efficiency of the image processing is low.

SUMMARY OF THE INVENTION The present invention has been made in consideration of the above problems, and has as its object to reduce the memory capacity in compression processing and to improve the efficiency of image processing in high-efficiency coding using DCT. It is another object of the present invention to provide an image processing method which can achieve the image reproducibility caused by decompressed data.

[0006]

Means for Solving the Problems] To achieve the above object, the present invention compresses the incoming Chikarade chromatography data, and the compressed
To extend the data, the same input data DCT (discrete cosine transform) in block units is calculated, and an image processing method for inverse DCT operation of the compression de <br/> over data, entering-force in data compression, reducing the number of bits by truncating a lower predetermined bits before Symbol DCT calculation data,
When Ru obtain a compressed data that by the data having a reduced number of the bit
Both when extending the pre-Symbol compressed data, prior <br/> Symbol lower predetermined bit truncation fixed data by sign bits, extendable and inverse DCT operation the compressed data is fixed the lower predetermined bits It is characterized by
You .

[0007]

According to the above method, the DCT operation is performed on the input data (eg, 8-bit image data),
The lower bits (for example, 4 bits) of the T data (for example, 12 bits) are truncated, and the remaining 8
An encoding process is performed on the bit data.

Therefore, the capacity of the memory used in the above-described image data compression processing can be reduced, and the processing can be easily performed, and the efficiency of the processing can be improved.

When the compressed data is decompressed and the image data is decoded, the compressed data is converted into 8 bits, and a sign bit is added as the lower 4 bits of the 8-bit data. .

That is, the input data of DCT and inverse DCT are signed numbers, and the most significant bit of the data is a sign bit. For example, a positive number is "0", and a negative number is "1". It is.

Therefore, the lower 4 bits of the positive data in the 8-bit compressed data are set to "0000",
The lower 4 bits of the negative data are set to “1111”,
The quantization levels of the data are made stepwise.

As is apparent from FIG. 5, the step-like quantization level is biased toward the reference "0" for the positive data and the negative data, that is, the inverse DCT
Input is biased to the middle level. Therefore, its inverse DCT
It can be said that the data is normal for the operation, and even if the result of the inverse DCT operation is A / D converted, the brightness and color difference of the image are not biased.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS In the image processing method of the present invention, when compressing input data (image data), the input image data is subjected to DCT (discrete cosine transform) operation. When the compressed data is expanded, the lower predetermined bits of the compressed data (the lower predetermined bits cut off) are fixed to predetermined data, and the lower predetermined bits are converted to the fixed compressed data. And performs an inverse DCT operation.

[0014] Hereinafter, to FIGS. 1 to 5 the image processing method will be described in detail with refer to the present invention. In FIG. 1, DCT
(Discrete Cosine Transform) 1 assumes that the input image data is 8 bits and the output data is 12 bits.

The input of the DCT1 and the inverse DCT2 is a signed number. For example, in the case of data of 0 to 255, the input image data (8 The sign bit (“0”, “1”) representing positive or negative is added to the most significant bit (eighth bit) of the bit.

Then, the 8-bit input image data is subjected to a DCT operation, the result of the DCT operation is output in 12 bits, and the input image data is compressed.

Then, as shown in the schematic diagram of the data in FIG. 2, the 12-bit data (D0, D0,
.., D11) are positive and negative. For example, 0 to 2047 on the positive side and −
Data of 1 to -2048 are obtained. The above 12
The most significant bit D11 of the bit indicates positive or negative.

Subsequently, the lower 4 bits (D0, D1, D2, D3) of the sorted data are discarded, and FIG.
, +20 to +15, +16 to +31,..., +20 of the DCT-calculated positive data.
Data of 32 to +2047 are set to the same value, and similarly, of the negative data, -2048 to -2
033, ..., -32 to -17, and -16 to -1 have the same value.

When processing such as run-length encoding is performed on the remaining 8 bits of data obtained by truncating the above 4 bits, the amount of memory is small because the number of bits is smaller by 4 bits. It becomes easier and the efficiency of the processing can be improved. On the other hand, when the compressed data is decompressed and decoded, fixed data is added as the lower 4 bits D0, D1, D2 and D3 of the 8-bit compressed data. "Or" 1111 ".

As a result, as shown in FIG. 3, when the lower 4 bits of the compressed data are fixed to "0000", the quantization level of the input data of the inverse DCT 2 becomes stair-like.
Further, as shown in FIG. 4, when fixed to "1111" the lower 4 bits of the compressed data, quantization levels of the input data of the inverse DCT2 becomes the same stepwise.

When the lower 4 bits of each data are fixed to "0000", the quantization level of positive data is biased toward "0", but the quantization level of the negative data is more negative. Biased in the direction. In addition, the lower 4
When the bit is fixed to “1111”, the quantization level of positive data is more biased toward the positive direction, and the quantization level of negative data is biased toward “0”.

Therefore, the compressed positive data (8
Bit), the sign bit “0” indicating positive is used, and the sign bit “1” indicating positive is used as the lower 4 bits of the negative data (8 bits). Are set to "0000", and the lower 4 bits of the negative data are set to "1111".

Then, the quantization level of the compressed data has a step-like shape as described above, but the quantization level of each positive data and the quantization level of the negative data are respectively shifted in the direction of “0” ( (Shown in FIG. 5).

Therefore, since the input of the inverse DCT 2 is biased to the intermediate level, that is, the input data of the inverse DCT 2 is normal, the inverse DCT operation is performed to expand the data, and the expanded data is converted to D / D A conversion,
The luminance and color difference of the image are not biased to one side, and the deterioration of the reproduced image is suppressed.

[0025]

As described above, according to the present invention, in an image processing method in which input image data is subjected to DCT operation and compressed, and compressed data is subjected to inverse DCT and decompression, a predetermined lower order of the DCT operated data is provided. Bits are truncated, the number of bits of the DCT-calculated data is reduced, and compressed data is obtained by the reduced data. When decompressing the compressed data, the truncated lower predetermined bits are fixed to predetermined data, In addition, since the quantization levels of the positive data and the negative data are each biased to the reference “0”, and the data of these quantization levels can be subjected to the inverse DCT operation, the compression of the input image data requires
Since the number of bits of the data subjected to the T operation is reduced, the capacity of the memory in the compression processing may be reduced, and the encoding processing and the like may be easily performed, and the efficiency of the processing may be improved.

In the image processing of the present invention, when decompressing the compressed data, the truncated lower predetermined bits of the DCT-calculated data are fixed with sign bits, and the positive data and the negative data of the data are fixed. Are biased toward the “0” side, that is, since the input of the inverse DCT is biased to an intermediate level, the result of the inverse DCT operation is D / A converted. However, the luminance and color difference of the image are not biased to one side, that is, the deterioration of the reproduced image is suppressed.

[Brief description of the drawings]

FIG. 1 is a schematic block diagram illustrating an embodiment of the present invention, to which an image processing method is applied.

FIG. 2 is a schematic diagram of data for explaining an image processing method of the present invention.

FIG. 3 is a graph of the data shown in FIG. 2;

FIG. 4 is a graph of the data shown in FIG. 2;

FIG. 5 is a graph of the data shown in FIG. 2;

[Explanation of symbols]

 1 DCT (Discrete Cosine Transform) 2 Inverse DCT

Claims (2)

(57) [Claims] To compress input data and decompress the compressed data, the input data is divided into blocks by D.
CT (Discrete Cosine Transform) calculated, and an image processing method for inverse DCT operation of the compressed data, when compressing the input data, the number of bits by truncating a lower predetermined bits of the data the DCT calculation reduce, <br/> compressed data by the data having a reduced number of the bits obtained Rutotomoni, when decompressing data the compressed fixes the lower predetermined bits truncated data by the sign bit, the lower position predetermined bits Inverse DCT of compressed data with
An image processing method characterized in that it can be decompressed by calculation. 2. The image processing method according to claim 1, wherein the data to be compressed is 8 bits, and the lower predetermined bits are at least 4 bits, and all of the 4 bits are the same data as the sign bit.