| 摘要 |
An image processing device includes a determination unit configured to determine a correlation value between a decoded frame and a reference frame and the number of iterations indicating the number of times signal separation calculation based on a TV method is iteratively performed on the decoded frame, based on a result of a motion search process, a generation unit configured to generate, as an initial gradient signal value, a zero vector or a gradient signal value obtained based on a motion vector determined in the motion search process, according to the correlation value, and a separation unit configured to iteratively perform the signal separation calculation on the decoded frame the number of iterations, using the initial gradient signal value as an initial value, and separate the decoded frame into a skeleton image signal and a texture image signal. |
