- [Home]
- [Research achievement]
- [Research achievement detail]
Title | A Note on Support Recovery of Sparse Signals using Belief Propagation (in Japanese) |
---|---|
Authors | Shunsuke Horii 、Toshiyasu Matsushima 、Shigeichi Hirasawa |
Released Year | 2016 |
Format | Conference |
Category | Others |
Jounal Name | |
Jounal Page | vol.IEICE-116, no.33(IT), no.34(EMM) , pp.19-24 |
Published Year | 2016 |
Published Month | 5 |
Abstract (English) |
In this paper, we propose a new support recovery algorithm from noisy measurements based on the belief propagation (BP) algorithm. We focus on the problem to recover the support of sparse signals rather than the problem to estimate sparse signals themselves. Previously, the idea to apply the BP algorithm to the support recovery problem was suggested. However, our proposed algorithm is different from the existing one in that our algorithm exchanges only discrete messages while the existing one exchanges continuous messages. As a result, we need not to rely on numerical integration or approximation method to compute messages and we can compute them analytically. Further, we also provide computationally efficient algorithm for binary sparse measurement matrices. We compare the performance of the proposed algorithm to the Sparse Matching Pursuit (SMP) via numerical simulations. |
Note (English) |
1 |
Manuscript | |
Presentation |