| 摘要 |
<p>Moving picture decoding arrangements where the motion vector for the target block of decoding is calculated on a decoding side. A reference frame is selected dependent on whether candidate blocks adjacent to the target block of decoding have a motion vector for forward prediction or backward prediction. In calculating the motion vector, a median of motion vectors of the adjacent blocks is set as the motion vector. Decoding of a moving picture is performed by using information of the selected reference frame and the calculated motion vector.</p> |
