| 摘要 |
<p>PURPOSE: To adaptively reduce the number of quantization data contained in one block and to perform variable length encoding in the compression-encoding of still pictures. CONSTITUTION: The pictures of one frame are inputted to a blocking part 11 and divided into the plural blocks of an (n)×(n) picture unit. Next, DCT transformation is performed in an ADCT transformation part 12 and a DC coefficient and an AC coefficient are obtained. The transformation coefficients are quantized in a quantization part 13. Further, in a quantization data transformation part 14, in a signal series obtained when the AC coefficient is zigzag- scanned, invalid data whose value is '0' and valid data whose value is not '0' are detected. Then, the run length and frequency of the invalid data are discriminated and the valid data thereafter are abandoned on the basis of the run length and frequency of the invalid data positioned before the valid data at a final position. The ones whose number pieces of data is reduced in such a manner are supplied to an entropy encoding part 15 and encoded.</p> |
