Efficient processing of compressed video

Compressed digital video is becoming prevalent for video storage and communication. It is often desirable to perform various video processing tasks on compressed video streams, however, operations once considered simple, such as splicing and reverse play, are much more complex in the context of compressed video. The conceptually-simple approach of completely decoding the video into the pixel domain, performing the processing, and re-encoding the entire video sequence is often inefficient and impractical. This paper investigates the problem of transcoding compressed video data from one standards-compliant data stream to another. The goal is to develop video transcoding algorithms that achieve high performance with computational efficiency. In particular, this paper presents efficient algorithms for splicing and reverse play of compressed video streams.