Approximate message passing-based compressed sensing reconstruction with generalized elastic net prior

作者：

Highlights：

• We study compressed sensing when an initial estimation of the signal is available.

• We propose an approximate message passing (AMP)-based solution.

• Parameter selection, state evolution, and noise-sensitivity analysis are presented.

• A practical parameterless version of the proposed method is also developed.

• Simulation results demonstrate the efficiency of the proposed schemes.

摘要

Highlights•We study compressed sensing when an initial estimation of the signal is available.•We propose an approximate message passing (AMP)-based solution.•Parameter selection, state evolution, and noise-sensitivity analysis are presented.•A practical parameterless version of the proposed method is also developed.•Simulation results demonstrate the efficiency of the proposed schemes.

论文关键词：Compressed sensing,Approximate message passing,Elastic net prior,State evolution,Phase transition

论文评审过程：Received 16 February 2015, Revised 29 June 2015, Accepted 29 June 2015, Available online 15 July 2015, Version of Record 3 August 2015.

论文官网地址：https://doi.org/10.1016/j.image.2015.06.011