| 摘要 |
PCT No. PCT/JP92/00910 Sec. 371 Date Apr. 19, 1993 Sec. 102(e) Date Apr. 19, 1993 PCT Filed Jul. 16, 1992 PCT Pub. No. WO93/02528 PCT Pub. Date Feb. 4, 1993In a method of compressing a motion picture signal divided into blocks of picture data and in which each of the blocks is selectively compressed by discrete cosine transform (DCT) coding or by intra-block predictive coding which effects non-transform coding (NTC); the intra-block predictive coding is effected by calculating a mean value of the picture data in a flat region of each block, determining a representation value of each block on the basis of the mean value thereof, determining the differences between the representation value for each block and the values of the pixels of picture data in the respective block, quantizing the resulting difference values with a selected quantizing width for generating quantizing coefficients, and variable length coding such quantizing coefficients for each block, the representation value of the respective block and the respective quantizing width. In decoding a bit stream of coded motion picture data, the bit stream is variable length decoded to generate the quantization coefficients for each block, the representation value determined on the basis of the mean value of a flat region in the respective block and information as to the quantizing width for the respective block, and then the generated quantization coefficients are decoded on the basis of the generated representation value and the information as to the quantizing width.
|
