Image compression apparatus using a lossy compression method

Claim 2

An image compression apparatus comprising: means for dividing an original image data into a plurality of image data blocks each of which is constituted by a plurality of pixels; preprocessing means for calculating a mean value of pixels of each of the plurality of image data blocks obtained by said dividing means, and for subtracting the mean value of pixels of each block for the pixels located at corresponding block, to output a plurality of preprocessed outputs each corresponding to a difference between the mean value and each of the pixels; and compression circuit means for orthoqonally-transforming, for each block, the preprocessed outputs obtained from said preprocessing means and outputting a plurality of transformed outputs corresponding to the image data blocks and for quantizing each of the transformed outputs to compress the image data, said compression circuit means including a first table for storing bit-number data representing the number of bits allocated in accordance with the frequency component of each of the orthoqonally-transformed image data blocks obtained by said orthoqonally-transforming means, a second table for storing data for limiting a quantization band to decrease a quantization band width as the number of bits allocated to the image data by said first table is reduced, and quantization means for quantizing the image data blocks in accordance with the data stored in said first and second tables.

Claim 3

An apparatus according to claim 1, wherein said compression circuit means comprises division means for dividing the preprocessed outputs obtained from said preprocessing means into a plurality of image data blocks, first orthogonal transform means for orthogonal-transforming the data blocks obtained by said division means in units of blocks and generating the transformed outputs, means for reconstructing the transformed outputs obtained by said first orthogonal transform means in units of identical frequency components and generating reconstructed outputs, second orthogonal transform means for orthogonally-transforming the reconstructed outputs obtained by said reconstructing means and generating second orthogonally-transformed outputs, and quantization means for quantizing the second orthogonally-transformed outputs obtained by said second orthogonal transform means.

Claim 4

An apparatus according to claim 2, wherein said division means comprises memory means for storing image data, and means for addressing said memory means in order to read out the preprocessed outputs in units of blocks.

Claim 5

An apparatus according to claim 2, wherein said reconstructing means comprises memory means for storing the transformed outputs from said first orthogonal transform means, and addressing means for outputting address data to said memory means so as to read out the transformed outputs in units of identical frequency components from said memory means.

Claim 6

An apparatus according to claim 2, wherein said second orthogonal transform means comprises orthogonal transform means for implementing one-dimensional orthogonal transform of the reconstructed outputs from said reconstructing means.

Claim 7

An apparatus according to claim 2, further comprising selection means for discriminating the reconstructed outputs from said reconstructing means to be components requiring orthogonal transform and components requiring no orthogonal transform, and selectively supplying the reconstruction outputs to said second orthogonal transform means and said quantization means in accordance with the discrimination result.

Claim 8

An image compression apparatus comprising: means for dividing an original image data into a plurality of image data blocks, each of which is constituted by a plurality of pixels; preprocessing means for calculating a set of pixel mean values, each being a mean value of pixels which are located at identical coordinates within two-dimensional spatial coordinate domains, and the plurality of image data blocks obtained by said dividing means, and for subtracting the mean value of pixels of each block from the pixels located at a corresponding block to output preprocessed outputs each corresponding to a difference between each of the pixels and the means value; and compression circuit means for orthoqonally-transforming for each block, the preprocessed outputs obtained from said preprocessing means to output a plurality of transformed outputs corresponding to the image data blocks and for quantizing each of the transformed outputs to compress the image data, said compression circuit means including a first table for storing bit-number data representing the number of bits allocated in accordance with a frequency component of each of the orthoqonally-transformed image data blocks obtained by said compression circuit means, a second table for storing data for limiting a quantization band to decrease a quantization band width as the number of bits allocated to the image data by said first table is reduced, and quantization means for quantizing the image data blocks in accordance with the data stored in said first and second tables.

Claim 9

An image compression apparatus comprising: means for dividing an original image data into a plurality of image data blocks, each of which is constituted by a plurality of pixels; preprocessing means for calculating a set of pixel mean values, each being a mean value of pixels which are located at identical coordinates within two-dimensional spatial coordinate domains, and the plurality of image data blocks obtained by said dividing means, and for subtracting the mean value of pixels of each block from the pixels located at a corresponding block to output preprocessed outputs each corresponding to a difference between each of the pixels and the means value; and compression circuit means for orthoqonally-transforming for each block, the preprocessed outputs obtained from said preprocessing means to output a plurality of transformed outputs corresponding to the image data blocks and for quantizing each of the transformed outputs to compress the image data, said compression circuit means including a first table for storing bit-number data representing the number of bits allocated in accordance with a frequency component of each of the orthoqonally-transformed image data blocks obtained by said compression circuit means, a second table for storing data for limiting a quantization band to decrease a quantization band width as the number of bits allocated to the image data by said first table is reduced, and quantization means for quantizing the image data blocks in accordance with the data stored in said first and second tables.

Claim 10

An apparatus according to claim 7, wherein said compression circuit means comprises division means for dividing the preprocessed outputs obtained from said preprocessing means into a plurality of data blocks, first orthogonal transform means for orthogonally-transforming the data blocks obtained by said division means in units of blocks and generating the transformed outputs, means for reconstructing the transformed outputs obtained by said first orthogonal transform means in units of identical frequency components and generating reconstructed outputs, second orthogonal transform means for orthogonally-transforming the reconstructing outputs obtained by said reconstruction means and generating second orthogonally-transformed outputs, and quantization means for quantizing the second orthogonally-transformed outputs obtained by said second orthogonal transform means.

Claim 11

An apparatus according to claim 8, wherein said division means comprises memory means for storing image data, and means for addressing said memory means in order to read out the preprocessed outputs in units of blocks.

Claim 12

An apparatus according to claim 8, wherein said reconstructing means comprises memory means for storing the transformed outputs from said first orthogonal transform means, and addressing means for outputting address data to said memory means so as to read out the transformed outputs in units of identical frequency components from said memory means.

Claim 13

An apparatus according to claim 8, wherein said second orthogonal transform means comprises orthogonal transform means for implementing one-dimensional orthogonal transform of the reconstructed outputs from said reconstructing means.

Claim 14

An apparatus according to claim 8, further comprising selection means for discriminating the reconstructed outputs from said reconstructing means to be components requiring orthogonal transform and components requiring no orthogonal transform, and selectively supplying the reconstruction outputs to said second orthogonal transform means and said quantization means in accordance with the discrimination result.

Claim 15

An image compression apparatus comprising: means for dividing an original image data into a plurality of image data blocks, each of which is constituted by a plurality of pixels; orthogonal transformation means for orthogonally-transforming the image data blocks output from said dividing means and outputting a plurality of frequency components including a plurality of identical frequency components and corresponding to the pixels; a first table for storing bit-number data representing the number of bits allocated in accordance with the frequency component of each of the orthogonally-transformed image data blocks obtained by said orthogonal transformation means; a second table for storing data for limiting a quantization band to decrease a quantization band width as the number of bits allocated to the image data by said first table is reduced; and quantization means for quantizing the image data blocks in accordance with the data stored in said first and second tables.

Claim 16

An image compression apparatus comprising: means for dividing an original image data into a plurality of image data blocks, each of which is constituted by a plurality of pixels; orthogonal transformation means for orthogonally-transforming the image data blocks output from said dividing means and outputting a plurality of frequency components including a plurality of identical frequency components and corresponding to the pixels; a first table for storing bit-number data representing the number of bits allocated in accordance with the frequency component of each of the orthogonally-transformed image data blocks obtained by said orthogonal transformation means; a second table for storing data for limiting a quantization band to decrease a quantization band width as the number of bits allocated to the image data by said first table is reduced; and quantization means for quantizing the image data blocks in accordance with the data stored in said first and second tables.