语音自适应压缩采样匹配追踪重构算法  

A Speech Reconstruction Algorithm Using Adaptive Compressive Sampling Matching Pursuit

作　　者：石磊[1] 

机构地区：[1]南京邮电大学宽带无线通信与传感网技术教育部重点实验室,江苏南京210003

出　　处：《南京邮电大学学报（自然科学版）》2013年第1期16-22,共7页Journal of Nanjing University of Posts and Telecommunications：Natural Science Edition

基　　金：国家自然科学基金(60971129)资助项目

摘　　要：基于小波分解下的语音压缩编码与重构框架,研究分析了含噪情况下贪婪算法的重构性能和抗噪性能,提出了一种改进的自适应压缩采样匹配追踪算法(ACoSaMP)。该算法可在稀疏度未知的情况下,通过设置可变步长分阶段实现对稀疏度的逼近。同时,在每次迭代过程中,用最小二乘法对残差信号进行估计,代替传统CoSaMP算法对整个信号的估计。最后用小波去噪法对合成语音进行处理。实验结果表明:不同压缩比下,该算法的主客观重构效果均优于现有同类算法,对噪声有较强的鲁棒性。This article ation based on speech analyzes greedy algorithmg reconstruction and anti-noise performances in noise situ- compression coding and reconstruction framework via wavelet decomposition. We propose an improved adaptive compressive sampling matching pursuit algorithm （ACoSaMP）. The algo- rithm can approximate the circumstances. Meanwhile sparsity in different phases by setting the variable step size in sparsity unknown , in every iteration, the residual signal is estimated using the least squares meth- od,instead of the traditional CoSaMP algorithm. synthesize speech. Experimental results show the outperform existing similar algorithms in different ness to noise. Finally, the wavelet de-noising method is employed to subjective and objective performances of the algorithm compression ratios and our algorithm has strong robut-

关 键 词：压缩感知 重构算法 自适应压缩采样 匹配追踪 小波去噪 

分 类 号：TN912[电子电信—通信与信息系统]