摘要
给出了求解块稀疏压缩感知的光滑加权块l1算法的理论分析,并通过数值仿真实验与3类具有代表性的l1-magic算法、SL0算法和FPC-AS算法进行了对比.实验结果表明,基于块结构的光滑加权块l1算法能更加有效地处理块稀疏信号.
Convergence analysis of the smooth weighted block-l1 algorithm for solving block-sparse compressed sensing is presented in this paper,and the smooth weighted block-l1 algorithm is compared with three representative algorithms,including l1-magic algorithm,SL0 algorithm and FPC-AS algorithm by numerical simulation experiments.The experimental results indicate that the smooth weighted block-l1 algorithm based on the block structure is better than traditional non-block algorithms in processing blocksparse signals.
出处
《西南大学学报(自然科学版)》
CAS
CSCD
北大核心
2014年第5期72-77,共6页
Journal of Southwest University(Natural Science Edition)
基金
国家自然科学基金资助项目(61273020
11001227)
关键词
压缩感知
块稀疏
光滑加权块l1算法
收敛性
compressed sensing
block sparsity
smooth weighted block-l1 algorithm
convergence