In visual communication systems, VLC(variable length code) is widely adopted for data compression. The VLC decoder could be realized by the existing serial or parallel processing methods. But these methods require high clock-rate or too much memories. These methods for the VLC decoding only concern the systems in view point of hardware operation.
In this paper, a VLC decoding method based on codeword paritioning is presented. We approached it from the properties of the uniquely decodable and prefix-free code. We analyze the properties of VLC, introduced the concept of codeword parititioning based on these properties. And we showed that the codeword paritioning could be applied to any set of optimum codewords of any information sources. We also present an optimum codeword partitioning procedure which is efficient for reducing hardware complexity.
In the experiment, we applied our scheme to the VLC of MPEG-1 AC coefficient. So we get the processing speed faster than serial processing method and remarkably reduced memory size ( about 0.8% - 0.5%) compared with the conventional parallel processing decoder.
