摘要 |
A motion vector determining method is provided to save additional usage of a memory by defining a list 1 reference picture for a direct mode as a lately decoded picture, and improve possibility that a direct mode is selected as a prediction mode of a macro block. If a co-located block located in a reference picture has list 1 and list 0 motion vectors, the list 0 motion vector of the co-located block is selected as a motion vector for inducing a motion vector of a B(Bi-predictive block). The selected motion vector is scaled on the basis of a temporal distance between the reference picture and a current picture. At least one motion vector of the B block is induced by applying bit operation to the selected motion vector. |