| 摘要 |
<P>PROBLEM TO BE SOLVED: To improve the accuracy of classification and to improve the quality of a converted image signal. <P>SOLUTION: A class generator 121 extracts edge components of reference data and residue data, detects motion vector information from the edge components as information for showing the inaccuracy of motion-compensating vector information used for decoding, and generates a class code CL0 based on the motion vector information. Using the class code CL0, etc., a classifier 123 obtains a class code CL showing the class to which pixel data at the interested position in an image signal Vb belong. A tape selector circuit 122 electively fetches pixel data xi at a predictive tape. Using the pixel data xi, and coefficient data Wi read in correspondence with the class code CL from a coefficient memory 124, an operation circuit 125 obtains pixel data y at the targeted position in the image signal Vb based on an estimation expression. <P>COPYRIGHT: (C)2004,JPO |
