论文部分内容阅读
文章提出了一种最大概率匹配的矢量量化编码算法，它为码书中的每一码字增加一个计数器，统计在编码图象时每个码字的出现的频数，并进行排序；在量化矢量时，根据当前码字出现频数大小依次选择侯选码字，即频数大的码字优先选为候选码字。该算法可以和已有的预测法结合，形成预测加最大概率匹配的联合矢量量化编码算法。实验表明，联合算法的效率较高，在最初几次的搜索中就能以较高的命中率命中最佳匹配码字。
This paper presents a vector quantization coding algorithm with maximum probability matching. It adds a counter for each codeword in the codebook, and counts the frequency of occurrence of each codeword when coding the image. In the quantization vector , The candidate codewords are selected according to the frequency of occurrence of the current codeword, that is, the codeword with the largest frequency is preferably the candidate codeword. The algorithm can be combined with the existing prediction method to form a joint vector quantization coding algorithm with maximum probability matching prediction plus. Experiments show that the joint algorithm is more efficient and can hit the best matching codeword with a higher hit rate in the first few searches.