Image signal encoding method by orthogonal transformation

Claim 1

A method of image signal encoding by orthogonal transformation wherein the orthogonal transformation is conducted on each block of two-dimensional image signals on the basis of a predetermined orthogonal function, and the transformed signals obtained by the orthogonal transformation are put side by side in the order of sequency of said orthogonal function in the block and are respectively encoded with an intrinsic code length, wherein the improvement comprises the steps of: (i) predicting code length allocation suitable for a block adjacent to said block, which has already been subjected to said encoding, on the basis of encoded signals and code length allocation in said block already subjected to said encoding, and (ii) encoding respective transformed signals in said adjacent block in accordance with said predicted code length allocation.

Claim 2

A method of image signal encoding by orthogonal transformation wherein the orthogonal transformation is conducted on each block of two-dimensional image signals on the basis of a predetermined orthogonal function, and the transformed signals obtained by the orthogonal transformation are put side by side in the order of sequency of said orthogonal function in the block and are respectively encoded with an intrinsic code length, wherein the improvement comprises the steps of: (i) predicting code length allocation suitable for a block adjacent to said block, which has already been subjected to said encoding, on the basis of encoded signals and code length allocation in said block already subjected to said encoding, and (ii) encoding respective transformed signals in said adjacent block in accordance with said predicted code length allocation.

Claim 4

A method of image signal encoding by orthogonal transformation wherein the orthogonal transformation is conducted on each block of two-dimensional image signals on the basis of a predetermined orthogonal function, and the transformed signals obtained by the orthogonal transformation are put side by side in the order of sequency of said orthogonal function in the block and are respectively encoded with an intrinsic code length, wherein the improvement comprises the steps of: (i) extracting said transformed signals; assembling said extracted signals having the same sequency in each of the respective blocks, into a separate group for each different sequency; and calculating a standard deviation of said transformed signal thus extracted, or a mean value of absolute values of said transformed signals thus extracted, for each said different sequency; and (ii) adjusting said code length for said transformed signal of each sequency on the basis of said standard deviation or said means value.

Claim 5

A method of image signal encoding by orthogonal transformation wherein the orthogonal transformation is conducted on each block of two-dimensional image signals on the basis of a predetermined orthogonal function, and the transformed signals obtained by the orthogonal transformation are put side by side in the order of sequency of said orthogonal function in the block and are respectively encoded with an intrinsic code length, wherein the improvement comprises the steps of: (i) extracting said transformed signals; assembling said extracted signals having the same sequency in each of the respective blocks, into a separate group for each different sequency; and calculating a standard deviation of said transformed signal thus extracted, or a mean value of absolute values of said transformed signals thus extracted, for each said different sequency; and (ii) adjusting said code length for said transformed signal of each sequency on the basis of said standard deviation or said means value.

Claim 6

A method of image signal encoding by orthogonal transformation wherein the orthogonal transformation is conducted on each block of two-dimensional image signals on the basis of a predetermined orthogonal function, and the transformed signals obtained by the orthogonal transformation are put side by side in the order of sequency of said orthogonal function in the block and are respectively encoded with an intrinsic code length, wherein the improvemet comprises the steps of: (i) extracting said transformed signals assembling said extracted signals having the same sequency in each of the respective blocks, into a separate group for each different sequency, and calculating a standard deviation of said transformed signal thus extracted, or a mean value of absolute values of said transformed signals thus extracted, for each said different sequency, and (ii) adjusting said code length for said transformed signal of each sequency on the basis of said standard deviation or said mean value; and wherein processing for adjusting said code length on the basis of said standard deviation or said mean value is conducted for the sequency corresponding to a spatial frequency not lower than a predetermined value, and said code length for the sequency corresponding to a spatial frequency lower than said predetermined value is adjusted to a length necessary for expressing the maximum transformed signal value which may appear in each sequency.

Claim 7

A method of image signal encoding by orthogonal transformation wherein the orthogonal transformation is conducted on each block of two-dimensional image signals on the basis of a predetermined orthogonal function, and the transformed signals obtained by the orthogonal transformation are put side by side in the order of sequency of said orthogonal function in the block and are respectively encoded with an intrinsic code length, wherein the improvemet comprises the steps of: (i) extracting said transformed signals assembling said extracted signals having the same sequency in each of the respective blocks, into a separate group for each different sequency, and calculating a standard deviation of said transformed signal thus extracted, or a mean value of absolute values of said transformed signals thus extracted, for each said different sequency, and (ii) adjusting said code length for said transformed signal of each sequency on the basis of said standard deviation or said mean value; and wherein processing for adjusting said code length on the basis of said standard deviation or said mean value is conducted for the sequency corresponding to a spatial frequency not lower than a predetermined value, and said code length for the sequency corresponding to a spatial frequency lower than said predetermined value is adjusted to a length necessary for expressing the maximum transformed signal value which may appear in each sequency.