Learning a compressive sensing matrix with structural constraints via maximum mean discrepancy optimization

Signal Processing（2022）

引用 4|浏览20

暂无评分

摘要

•Design of structured (example: constant modulus entries) compressive sensing matrices.•Enforcing a restricted isometry property formulated as distribution matching problem.•Distribution matching measured via maximum mean discrepancy and solved via learning.•Optimized matrix can outperform random matrices in numerical experiments.

查看译文

关键词

Compressive sensing,Machine learning,Maximum mean discrepancy,Restricted isometry property,Sparse channel estimation

AI 理解论文

溯源树

样例

生成溯源树，研究论文发展脉络

Chat Paper

正在生成论文摘要