On selection of search space dimension in compressive sampling matching pursuit
2012 (English)In: TENCON 2012 - 2012 IEEE Region 10 Conference, IEEE , 2012, 6412345- p.Conference paper (Refereed)
Compressive Sampling Matching Pursuit (CoSaMP) is one of the popular greedy methods in the emerging field of Compressed Sensing (CS). In addition to the appealing empirical performance, CoSaMP has also splendid theoretical guarantees for convergence. In this paper, we propose a modification in CoSaMP to adaptively choose the dimension of search space in each iteration, using a threshold based approach. Using Monte Carlo simulations, we show that this modification improves the reconstruction capability of the CoSaMP algorithm in clean as well as noisy measurement cases. From empirical observations, we also propose an optimum value for the threshold to use in applications.
Place, publisher, year, edition, pages
IEEE , 2012. 6412345- p.
Compressed sensing, Greedy Pursuit Algorithms, Sparse Recovery
Engineering and Technology
IdentifiersURN: urn:nbn:se:kth:diva-118766DOI: 10.1109/TENCON.2012.6412345ScopusID: 2-s2.0-84873975479ISBN: 978-146734822-5OAI: oai:DiVA.org:kth-118766DiVA: diva2:608124
2012 IEEE Region 10 Conference: Sustainable Development Through Humanitarian Technology, TENCON 2012, 19 November 2012 through 22 November 2012, Cebu
QC 201302262013-02-262013-02-262013-02-26Bibliographically approved