An improved group-based optimizing of CAVLC decoder

Variable length code is a general lossless entropy coding method that is widely used in video compress standards. Based on analyzing R. Hashemian's fixed-length group decoding methods and the structure and statistic principles of CAVLC code words, this paper puts forward a kind of decoding method by dividing the CAVLC code words into prefix and suffix. The method has been proved to save 62.3% memory and decrease the needed decoding speed more than R. Hashemian's fixed-length group decoding methods.