基于概率补偿的无哈夫曼树变长压缩编码  

Variable-length Compressed Encoding Without Huffman Tree Based on the Probability Compensation

作　　者：杨多星[1] 刘蕴红[1] 

机构地区：[1]大连理工大学电气工程学院,辽宁大连116023

出　　处：《微电子学与计算机》2011年第6期51-53,57,共4页Microelectronics & Computer

摘　　要：现在广泛使用的压缩编码方法都要通过哈夫曼树来实现,这样围绕着哈夫曼树就存在着许多运算过程.为了化简编码过程,提出了一种无需哈夫曼树就能实现的变长最佳编码方法,通过一个概率补偿的过程,可以直接得到所有信源的最佳码长.知道码长和概率后也无需通过哈夫曼树就可以确定最后的编码,并且可以证明结果满足变长最佳编码定理和前缀编码.经测试,该方法可以快速有效得到变长最佳编码,并简化了变长编码的运算存储过程.Nowadays most of the widely used compressions encoding methods are implemented by using the Huffman tree.There are many operational processes around the Huffman tree.One variable-length optimal encoding method is proposed to simplify the encoding process.The optimal code length can be get from the probability compensation process.The final coding can be determined without the Huffman tree after knowing the code length and probability,and the result meets the challenge of the variable-length optimal coding theorem and the prefix code.After testing,the method is proved to be able to get the variable-length optimal coding quickly and effectively.The calculation and stored procedures are simplified.

关 键 词：哈夫曼树 变长最佳编码 概率补偿 前缀编码 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]