| 摘要 |
PROBLEM TO BE SOLVED: To provide a method for preventing mismatching of a predicted image caused by a difference of arithmetic accuracy of a motion vector without mak ing arithmetic operations complicated with respect to a motion compensation system where there is no restriction that all pixels belonging to the same patch have to have a common motion vector and the horizontal/vertical components of the motion vector of the pixels can take a value other than an integer multi ple of the distance between adjacent pixels. SOLUTION: The horizontal/vertical components of the motion vector of each pixel used for composting a prediction image are specified to take values of an integer multiple of 1/d1, 1/d2 of respective distances between adjacent pixels (d1, d2 are positive integers). Thus, mismatching of the prediction image caused by the difference of arithmetic accuracy of a motion vector can be prevented with a simple arithmetic operation so as to simplify the processing. COPYRIGHT: (C)2003,JPO |
