Video signal encoding method for e.g. digital television systems

摘要

The current motion vector encoding method is based on a number of reference motion vectors, that represent the displacement between a search block in a current frame and a reference block in a previous frame. The method utilizes motion vectors which include both horizontal and vertical components. In a shape or shape-texture combined mode, valid reference motion vectors are found, the vectors having corresponding reference blocks that comprise a boundary of an object. Valid reference motion vectors are counted and a first selection signal is generated if the number of valid vectors is equal to zero, otherwise a second selection signal is produced. A predictor for the current motion vector among the reference vectors is then determined, where the predictor is selected first from the valid vector for shape if there is a valid vector for shape, and otherwise, selected from the valid vector for texture in shape-texture combined mode. To determine an optimum predictor the zero value is selected in response to the first selection signal or the predictor is selected in response to the second selection signal. As a result, of the optimum value, the difference between a component of the current motion vector and a first component of the optimum predictor is determined. The difference is also determined for the second component of the current motion vector and the second component of the optimum predictor, thereby generating encoded data of the current motion vector. An Independent claim is also included for an apparatus for encoding a current motion vector based on a plurality of reference motion vectors.