| Summary: | 碩士 === 國立中正大學 === 電機工程所 === 97 === This work develops a compromised scheme of reducing memory and complexity associated with H.264/AVC in which multiple reference frames are used in motion estimation/compensation for inter-predictive coding. While multiple reference frames are adopted to increase coding efficiency of inter-predictive coding, they also consume a lot of memory resource. If the memory space is insufficient, a video sequence can not be completely and correctly decoded to result in quality degradation. This work presents a novel scheme to store a portion of macro blocks as compressed data according to the constrained memory space. Additionally, this insufficient memory space is effectively used to minimize the complexity of decoding. The simulation results demonstrate that the proposed method can utilize less memory than the conventional ones in H.264/AVC. The overhead of the proposed method is slightly increased computational complexity. Therefore, the method proposed herein can be widely employed in various devices with insufficient memory to successfully decode H.264/AVC video sequences.
|
|---|
