| 摘要 |
<P>PROBLEM TO BE SOLVED: To decode encoded data into a high quality image and high quality sound. <P>SOLUTION: A mismatch detection part 1 judges the rightness of a DCT type included in the encoded data which is MPEG encoded and supplies mismatch information expressing its judgment result to a decoding processing part 2. The decoding processing part 2 constitutes a prediction tap in a structure suitable to the mismatch information by extracting an object to be defined as the prediction tap from the encoded data on the basis of the mismatch information. Furthermore, the decoding processing part 2 calculates a predicted value of high quality image data by a linear primary prediction operation by using the prediction tap and a tap coefficient calculated by preliminarily learning it. <P>COPYRIGHT: (C)2003,JPO |
