Proceedings of the 2017 5th International Conference on Frontiers of Manufacturing Science and Measuring Technology (FMSMT 2017)

Accelerated Hard Thresholding Algorithms for Sparse Recovery

Authors
Xueci Zhao, Peibing Du, Tao Sun, Lizhi Cheng
Corresponding Author
Xueci Zhao
Available Online April 2017.
DOI
https://doi.org/10.2991/fmsmt-17.2017.259How to use a DOI?
Keywords
Hard thresholding pursuit, restricted isometry property, compressive sensing, optimization methods.
Abstract
In linear observations, i.e., a system of linear equations ( ), the hard thresholding pursuit (HTP) is used to find a sparse signal. HTP is an iterative algorithm that has been found many applications in compressive sensing, due to its good recovery performance, which includes linear convergence speed, high recovery rate, and stability. In this paper, we further develop accelerated algorithms to deal with a linear least square (LLS) problem in each iteration. Theoretically, we prove that all these algorithms are convergent, provided that the sensing matrix has suitable restricted isometry property. Numerical experiments on sparse signal recovery demonstrate the efficiency of the proposed methods.
Open Access
This is an open access article distributed under the CC BY-NC license.

Download article (PDF)

Cite this article

TY  - CONF
AU  - Xueci Zhao
AU  - Peibing Du
AU  - Tao Sun
AU  - Lizhi Cheng
PY  - 2017/04
DA  - 2017/04
TI  - Accelerated Hard Thresholding Algorithms for Sparse Recovery
BT  - Proceedings of the 2017 5th International Conference on Frontiers of Manufacturing Science and Measuring Technology (FMSMT 2017)
PB  - Atlantis Press
SP  - 1322
EP  - 1328
SN  - 2352-5401
UR  - https://doi.org/10.2991/fmsmt-17.2017.259
DO  - https://doi.org/10.2991/fmsmt-17.2017.259
ID  - Zhao2017/04
ER  -