An Adaptive-ADMM Algorithm With Support and Signal Value Detection for Compressed Sensing
Authors:
- 8 February 2013
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 20 (4) , 315-318
Abstract
This letter presents a novel adaptive alternating direction method of multipliers with support/signal value detection for compressed sensing. The support/signal value detection in our algorithm can achieve an efficient reconstruction by leveraging more information that goes beyond simple sparsity. Especially for time-correlated signals in large-scale problems, our proposal performs better than conventional methods, since more accurate signal information could be estimated from prior knowledge during initialization. Simulation results show that our method can improve the average PSNR by 1.02-2.05 dB for undersampled video sequences.Keywords
This publication has 12 references indexed in Scilit:
- Exact Reconstruction Conditions for Regularized Modified Basis PursuitIEEE Transactions on Signal Processing, 2012
- Regularized Modified BPDN for Noisy Sparse Reconstruction With Partial Erroneous Support and Signal Value KnowledgeIEEE Transactions on Signal Processing, 2011
- Recovering Compressively Sampled Signals Using Partial Support InformationIEEE Transactions on Information Theory, 2011
- Distributed Optimization and Statistical Learning via the Alternating Direction Method of MultipliersFoundations and Trends in Machine Learning, 2011
- Modified-CS: Modifying Compressive Sensing for Problems With Partially Known SupportIEEE Transactions on Signal Processing, 2010
- Fast Image Recovery Using Variable Splitting and Constrained OptimizationIEEE Transactions on Image Processing, 2010
- Sparse Signal Reconstruction via Iterative Support DetectionSIAM Journal on Imaging Sciences, 2010
- A New Alternating Minimization Algorithm for Total Variation Image ReconstructionSIAM Journal on Imaging Sciences, 2008
- Robust uncertainty principles: exact signal reconstruction from highly incomplete frequency informationIEEE Transactions on Information Theory, 2006
- Convex Analysis1970