基于改进CoSaMP的图像重构算法

打开文本图片集

Image Construction Algorithm Based on Improved CoSaMP

Xu Xuejie, Zhao Qingping (HuaibeiNormal University,Huaibei235ooo,China)

【Abstract】 Aiming at the problem that Compressive Sampling Matching Pursuit algorithm (CoSaMP) needs prior knowledge of image sparsity,this paper proposesa sparsity-adaptive image reconstruction algorithm named Modified CoSaMP(MCSaMP).Firstly,amatching test strategy is employed toadaptivelyestimate theoptimal sparsitylevelof the image.Secondly,the atom selection criterion is refined based ona similarity metric.Finaly,an iteration stopping condition is established by analyzing the ℓ2 -norm relationship of residuals over the most recent three iterations. The simulations demonstrate thatthe proposed algorithm achieves superiorreconstruction performance without requiring prior sparsity information,outperforming conventional methods.

【Key words】compresivesensing;sparserepresentation; measurement matrix;reconstructionalgorithm;reconstruction performance

[中图分类号］TN911.7 [文献标识码]A [文章编号]1674-3229(2025)03-0048-07

0 引言

压缩感知理论[是传统采样定理的革新，相比于传统理论，它可以通过数量更小的采样值实现信号的重建。（剩余10262字）