| 摘要 |
PURPOSE:To improve the compression ratio of picture data by applying orthogonal transformation encoding to vector quantization when a vector corresponding to the group of the picture data in respective block at the minimum distortion is selected and vector-quantized. CONSTITUTION:An original picture data S to show one sheet of continuous tone pictures is made to pass through a block conversion circuit 10, and is converted into the data (x) of every rectangular block. The data (x) is inputted to a vector quantizer 11, and the vector, the vector element of which corresponds to the group of the original picture data at the minimum distortion, is obtained, and encoded data Dm, showing an identification number (m) to define this vector x(m), is outputted. The said data Dm is obtained for all the blocks, and is inputted to an orthogonal transformation circuit 13, and for instance, is two-dimentionally orthogonal-transformed. A transformed data Ym, obtained by this procedure, is arrayed in longitudinal and lateral directions in the order of the sequency of a function, which is the basis of the transformation, and is encoded 14. Thus, since a high energy component is given a long code length, and a low energy component is either given the short code length or cut off, the data is compressed. |
