Robust Signal Recovery With Highly Coherent Measurement Matrices
Authors:
- 8 November 2016
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 24 (3) , 304-308
Abstract
By embedding an ℓ p -norm noise constraint for p ≥ 2 into the recently emerged ℓ 1-2 method, in this letter, we study theoretically and numerically an ℓ 1-2 /ℓ p method for recovery of general noisy signals from highly coherent measurement matrices. In particular, the obtained theoretical results not only improve the condition deduced in [1] for Gaussian noisy signal recovery but also provide a new theoretical guarantee for generally nonGaussian noisy signal recovery. What is more, to better boost the recovery performance, a partial sum ℓ 1-2 /ℓ p method is also proposed latter. This improved method, together with the previous ℓ 1-2 /ℓ p method, becomes more competitive when compared with some of the state-of-the-art methods in recovering noisy signals from highly coherent measurement matrices.Keywords
Funding Information
- Natural Science Foundation of China (61273020, 61673015)
- Fundamental Research Funds for the Central Universities (XDJK2015A007)
- Natural Science Foundation of Chongqing Municipality (cstc2015gjhz40002)
This publication has 20 references indexed in Scilit:
- A null-space-based weightedl1minimization approach to compressed sensingInformation and Inference: A Journal of the IMA, 2016
- Outlier-Robust Greedy Pursuit Algorithms in $\ell _{p}$ -Space for Sparse ApproximationIEEE Transactions on Signal Processing, 2015
- Compressed Sensing with Non-Gaussian Noise and Partial Support InformationIEEE Signal Processing Letters, 2015
- Linear convergence of adaptively iterative thresholding algorithms for compressed sensingIEEE Transactions on Signal Processing, 2015
- Smoothed Low Rank and Sparse Matrix Recovery by Iteratively Reweighted Least Squares MinimizationIEEE Transactions on Image Processing, 2014
- $L_{1/2}$ Regularization: Convergence of Iterative Half Thresholding AlgorithmIEEE Transactions on Signal Processing, 2014
- Fast and Accurate Matrix Completion via Truncated Nuclear Norm RegularizationIEEE Transactions on Pattern Analysis and Machine Intelligence, 2012
- Dequantizing Compressed Sensing: When Oversampling and Non-Gaussian Constraints CombineIEEE Transactions on Information Theory, 2010
- Sparse Signal Reconstruction via Iterative Support DetectionSIAM Journal on Imaging Sciences, 2010
- Neuromagnetic source imaging with FOCUSS: a recursive weighted minimum norm algorithmElectroencephalography and Clinical Neurophysiology, 1995