Computational graceful degradation for video sequence decoding
暂无分享,去创建一个
The implementation of software based video decoders is an advantageous solution over traditional dedicated hardware real time systems. The main reason of this fact is the possibility of decoding several different video compression standards, at different levels and profiles, by just changing the software while using the same hardware platform, main memory, and interfaces. But, simultaneously guaranteeing real-time performance, needed when processing video/audio bitstreams, and an efficient use of the processing power, remains a very difficult task. We show the main reasons of such difficulties. We introduce the concept and the potential advantages of computational graceful degradation (CGD) for the efficient implementation of multimedia processors. We develop techniques for the application of CGD to the decoding of compressed video sequences. Finally, simulation results for the H.263 video compression standard are reported. The techniques can be virtually applied to any video compression standard and processor platform.
[1] Jane W.-S. Liu,et al. Scheduling Real-Time, Periodic Jobs Using Imprecise Results , 1987, RTSS.
[2] Wolfgang A. Halang,et al. Constructing Predictable Real Time Systems , 1991 .
[3] Jane W.-S. Liu,et al. Imprecise Results: Utilizing Partial Comptuations in Real-Time Systems , 1987, RTSS.