| 摘要 |
PURPOSE:To attain high-compressing rate encoding of picture date in a state where the gradation reducibility of pictures is excellent, by coding the picture data having gradation into 1st code at every block of a prescribed size and executing vector quantization to the picture blocks coded into the 1st codes. CONSTITUTION:A scanner 41 photoelectrically reads an original containing a picture to be coded with a CCD image sensor and outputs picture data V expressing the picture of the original. A dividing section 42 fetches the picture data V outputted from the scanner 41 and divides the picture data V into blocks of a prescribed size (for example, 4X4) composed of plural picture elements. A pre-processing section 43 finds the mean value, standard deviation, etc., of the picture data contained in each block outputted from the dividing section 42 so as to carry out a pre-process for vector quantization. A vector quantizing section 44 vector-quantizes each block data pre-processed at the section 43. Therefore, the compressing efficiency can be improved and, at the same time, increase in distortion to a reproduced picture can be suppressed. |
