摘要
由于以往对ＬＺＷ无损压缩算法的研究多是基于平稳遍历的信源这一假设，而实际的信源多为局部平稳，在此情况下，原ＬＺＷ算法存在的不足在很大程度上限制了其最终压缩比的提高．据此，提出了采用双串表自适应更新机制以取代原算法中的单一串表机制，能较好地适应一般信源的局部平稳性，实际的编码效率测试曲线显示出改进的算法较原算法在压缩比上有很大提高．在对四幅医学图象的无损压缩实验中，均取得了优于ＪＰＥＧ的压缩比．
The existing analyses of the LZW algorithm are mostly based on the hypothesis that the source is stationary ergodic. However, the actual sources are always local stationary,and in that case the drawbacks of the original algorithm will keep it from further increase in compression ratio. This paper adopts double string table adaptive mechanism instead of the original single string table,so it can be adjusted to the local stationarity. The measured compression ratio curves of the two algorithms show that the results of the modified one are much better than that of the original one. And in the implementation of the reversible compression of 4 medical images (512×512 8bit), the results of the presented algorithm are better than that of JPEG.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
1998年第9期110-113,共4页
Journal of Shanghai Jiaotong University
