| 摘要 |
<p>PURPOSE:To improve data compressibility, to decrease a calculating quantity so as to shorten a processing time and to reduce a circuit scale by dividing image data into the blocks of nXn picture elements, calculating a range and a median from the in-block maximum value and minimum value of vector respective elements and normalizing, quantizing and encoding an image vector. CONSTITUTION:The image data are divided 1 into the blocks of the nXn picture elements, the range and median are calculated 3, 4 from the in-block maximum value and minimum value of vector respective elements, the image vector is normalized 5, a vector quantization is executed 6, 7 to the normalized vector, simultaneously, the vector-quantized data are encoded 10, and the range and minimum value or the range and median are encoded. Thus, since it is unnecessary to prepare a code book for each block, a processing can be speeded up. Since an existing data compressor can be used, a device constitution can be simplified, the compressible and picture quality can be improved and, further, easily made into a high order.</p> |
