基于Hadamard变换和自适应顺序搜索的码字快速搜索算法  

A Fast Codeword Search Algorithm Based on Hadamard Transform and Adaptive Search Sequence

作　　者：乔阳[1] 潘志斌[1] 乔瑞萍[1] 李东平[1] 蔡骋[1] 

机构地区：[1]西安交通大学电子与信息工程学院,陕西西安710049

出　　处：《微电子学与计算机》2009年第1期118-121,共4页Microelectronics & Computer

基　　金：国家自然科学基金项目(60672054)

摘　　要：提出了一种Hadamard域中改进的快速码字搜索算法.在已离线按照码字第一维分量的大小进行了排序的码书中,首先找出与输入矢量第一维分量最接近的L个初始候选码字,求出对应的L个Chebyshev距离,接着按自适应的方法在这L个码字之外进行上下搜索,并用新找到的具有更小Chebyshev距离的码字来更新这L个候选码字,以便得到全体码书中L个具有最小Chebyshev距离的最终候选码字.最后用PDS算法在这L个最终候选码字中找出Euclidean距离最小的码字作为最佳匹配码字.实验表明文中算法相比本文算法在保证PSNR性能无任何下降的前提下,明显减少了算法的计算量,有效地提高了编码速度.An improved fast codeword search algorithm in Hadamard transform （HT） domain is proposed in this paper. First, the codewords in the HT domain are offline sorted according to their values of the first element, then a number of L codewords whose values of the first component are closest to a given input vector are selected, and the Chebyshev distances form the input vector to these selected L codewords are computed. Second, the codeword search starts around the initially selected L codewords adaptively, and they will be updated by the new codeword with less Chebyshev distance. Third, the best matched codeword can be obtained from finally selected codewords using the partial distortion search （PDS） algorithm. The simulation shows that the proposed algorithm in this paper can save a lot of computational costs meanwhile keep the same PSNR quality as that of the previous algrithm.

关 键 词：矢量量化 码字快速搜索 自适应搜索 HADAMARD变换 Chebyshev距离 Euclidean距离 

分 类 号：TN919.8[电子电信—通信与信息系统]