In one embodiment, the method includes selecting a list0 motion vector of the co-located image block in the reference picture for the bi-predictive image block as a motion vector for deriving motion vectors of the bi-predictive image block if the co-located image block has a list1 and a list0 motion vector; and scaling the selected motion vector of the co-located block in the reference picture based on a temporal distance between a current picture including the bi-predictive block and the reference picture. The motion vector of the bi-predictive block is derived based on the scaled motion vector of the co-located block.