Computer Vision And Pattern Analysis Laboratory Home Page  Home
People  People
Publications  Publications
Publications  Databases
Contact Information  Contact
Supported Research Projects  Supported Research Projects
Research Activites  Research Activites
Research Groups
SPIS - Signal Processing and Information Systems Lab.SPIS - Signal Processing and Information Systems Lab.
Medical Vision and Analysis Group  Medical Research Activities
Biometrics Research Group  Biometrics Research Group
SPIS - Signal Processing and Information Systems Lab.MISAM - Machine Intelligence for Speech Audio and Multimedia.
Knowledge Base
  Paper Library
Search-Based Methods for the Sparse Signal Recovery Problem in Compressed Sensing
Authors: Nazim Burak Karahanoglu
Published in: Sabanci University Research Database
Publication year: 2013
Abstract: The sparse signal recovery, which appears not only in compressed sensing but also in other related problems such as sparse overcomplete representations, denoising, sparse learning, etc. has drawn a large attraction in the last decade. The literature contains a vast number of recovery methods, which have been analysed in theoretical and empirical aspects.

This dissertation presents novel search-based sparse signal recovery methods. First, we discuss theoretical analysis of the orthogonal matching pursuit algorithm with more iterations than the number of nonzero elements of the underlying sparse signal. Second, best-first tree search is incorporated for sparse recovery by a novel method, whose tractability follows from the properly defined cost models and pruning techniques. The proposed method is evaluated by both theoretical and empirical analyses, which clearly emphasize the improvements in the recovery accuracy. Next, we introduce an itera-tive two stage thresholding algorithm, where the forward step adds a larger number of nonzero elements to the sparse representation than the backward one removes. The presented simulation results reveal not only the recovery abilities of the proposed method, but also illustrate optimal choices for the step sizes. Finally, we propose a new mixed integer linear programming formulation for sparse recovery. Due to the equivalency of this formulation to the original problem, the solution is guaranteed to be correct when it can be solved in reasonable time. The simulation results indicate that the solution can be found easily under some reasonable assumptions, especially for signals with constant amplitude nonzero elements.
  download full paper

Home Back