DocumentCode :
2366586
Title :
Turbo reconstruction of structured sparse signals
Author :
Schniter, Philip
Author_Institution :
Dept. ECE, Ohio State Univ., Columbus, OH, USA
fYear :
2010
fDate :
17-19 March 2010
Firstpage :
1
Lastpage :
6
Abstract :
This paper considers the reconstruction of structured-sparse signals from noisy linear observations. In particular, the support of the signal coefficients is parameterized by hidden binary pattern, and a structured probabilistic prior (e.g., Markov random chain/field/tree) is assumed on the pattern. Exact inference is discussed and an approximate inference scheme, based on loopy belief propagation (BP), is proposed. The proposed scheme iterates between exploitation of the observation-structure and exploitation of the pattern-structure, and is closely related to noncoherent turbo equalization, as used in digital communication receivers. An algorithm that exploits the observation structure is then detailed based on approximate message passing ideas. The application of EXIT charts is discussed, and empirical phase transition plots are calculated for Markov-chain structured sparsity.
Keywords :
Markov processes; equalisers; probability; signal reconstruction; EXIT charts; Markov chain structured sparsity; approximate inference; belief propagation; digital communication receivers; exact inference; hidden binary pattern; noisy linear observation; noncoherent turbo equalization; observation structure; pattern structure; signal coefficients; structured probabilistic prior; structured sparse signals; turbo reconstruction; Additive noise; Belief propagation; Digital communication; Gaussian noise; Heart; Inference algorithms; Message passing; Noise measurement; Reconstruction algorithms; Sparse matrices;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Information Sciences and Systems (CISS), 2010 44th Annual Conference on
Conference_Location :
Princeton, NJ
Print_ISBN :
978-1-4244-7416-5
Electronic_ISBN :
978-1-4244-7417-2
Type :
conf
DOI :
10.1109/CISS.2010.5464920
Filename :
5464920
Link To Document :
بازگشت