| 摘要 |
Embodiments of the present invention provide methods and apparatus for decoding digital video or motion picture signals with error concealment to prevent or minimize degradation of the picture quality. When corruption occurs, the method determines whether error concealment is needed according to a preset criterion. If so, motion vectors for the corrupt region are calculated using the motion vectors and rates of change of motion of current blocks in the current frame and/or old blocks in a previous frame according to a best matching block selection process. The method also determines if the "time" relative motion of content is small by defining a percentage use parameter. In some cases, once the error concealment is complete, the "weight" or visual sensitivity of the restored block is decreased in order to reduce boundary artifact effects. This is accomplished by a quantization mask employing the human visual system (HVS) concept.
|
