| 摘要 |
A method of image signal encoding comprises the steps of conducting two-dimensional orthogonal transformation on each block of two-dimensional image signals based on a predetermined orthogonal function, recognizing a region, where absolute values of almost all of the transformed signals obtained by the orthogonal transformation are smaller than a predetermined value, from a matrix of the transformed signals arranged in the order of sequency of the orthogonal function in the block, and sampling only the transformed signals outside of the recognized region. Or, characteristic values corresponding to absolute values of the transformed signals are added from the high frequency component side toward the low frequency component side, and only the transformed signals on the frequency side lower than the frequency corresponding to the transformed signal at which the sum obtained by the addition comes up to a predetermined value are sampled. The sampled signals are encoded respectively with an intrinsic code length.
|
