| 摘要 |
<p>PROBLEM TO BE SOLVED: To reduce the scale of a circuit by reducing the quantity of operation for motion vector detection by detecting a rough motion vector from the center of the entire search range of the motion vector detection, while using a content addressed memory(CAM), and detecting an optimum vector based on the result. SOLUTION: A data retrieval control part 103 inputs specified data in a block into a CAM 101 and finds a rough motion vector form an address provided as a result of calculation. Further, based on this rough motion vector, a data transfer control part 102 reads its peripheral area out of the CAM 101 and transfers it to an evaluated value operating part 104. At the evaluated value operating part 104, block matching is successively performed with respect to all the pixels between the data read out of the CAM 101 and the block, the respective matching results are compared, the optimum motion vector is determined, and that motion vector is reported to a motion compensation prediction part 211. Thus, the quantity of compilation for optimum motion vector detection can be reduced, and the circuit scale can be reduced.</p> |
