摘要 |
PROBLEM TO BE SOLVED: To improve a coding efficiency while suppressing a circuit scale from being increased. SOLUTION: An image memory section 1 stores an input image 101 one by one frame each for each prescribed time. A detection range division section 2 divides a detection range of a motion vector of the input image 101 into three detection ranges in response to a distance from a detection center, that is, a near detection range 103, a medium detection range 104 and a remote detection range 105. A high accuracy motion vector detection section 4 detects a motion vector with high accuracy. A medium accuracy motion vector detection section 5 detects a motion vector with medium accuracy. A low accuracy motion vector detection section 6 detects a motion vector with low accuracy. A motion vector selection section 7 selects a motion vector with least error among the three motion vectors. A prediction image generating section 8 generates a prediction image from the received image and the selected vector and provides an output of the predicted image. |