| 摘要 |
PURPOSE:To make faster an encoding processing without damaging a picture quality by encoding the orthogonal conversion of the data of a selected degenerated-block. CONSTITUTION:A picture data 1 are successively read out as picture elements with a block size 16X16 by a basic block reading part 2. For the read block data, for example, the dispersion of the luminance level of the picture elements in the block is calculated by a degenerated block classifying part 3, and based on the obtained luminance dispersion, thresholds Th1 and Th2 are provided, and the data are classified into three modes. The block data classified into modes are inputted to discrete cosine converting parts 41-43 corresponding to the picture elements with respective encoding data numbers 16X16, 8X8 and 4X4, and the conversion of the data is executed. The coefficients converted by the converting parts 41-43 are quantized by respective quantizing parts 51-53. A data selecting part 6 selects the encoding data obtained at every block, and stores or transmits the data. |
