| 摘要 |
<p>A method of predicting a block of a bi-predictive picture to be coded or decoded in direct mode is disclosed, the method comprising: a first step for calculating forward (MV f ) and backward (MV b ) motion vectors for the block ion the bi-predictive picture (B 5 ); a second step for obtaining a forward motion-compensated block (B f ) using the forward motion vector (MV f ) and a forward reference picture (P 1 ), and obtaining a backward motion-compensated block (B b ) using the backward motion vector (MV b ) and a backward reference picture (P 7 ); and a third step for predicting the block (B c '), wherein a first weight factor is applied to the forward motion-compensated block (B f ), and a second weight factor is applied to the backward motion-compensated block (B b ), wherein the first and second weight factors are based on respective temporal distances (TR B , TR D ).</p> |
