Efficient Sampling Set Selection for Bandlimited Graph Signals Using Graph Spectral Proxies

被引:253
作者
Anis, Aamir [1 ]
Gadde, Akshay [1 ]
Ortega, Antonio [1 ]
机构
[1] Univ So Calif, Dept Elect Engn, Los Angeles, CA 90089 USA
基金
美国国家科学基金会;
关键词
Graph signal processing; bandlimited graph signals; sampling set selection; experiment design; SPACES;
D O I
10.1109/TSP.2016.2546233
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
We study the problem of selecting the best sampling set for bandlimited reconstruction of signals on graphs. A frequency domain representation for graph signals can be defined using the eigenvectors and eigenvalues of variation operators that take into account the underlying graph connectivity. Smoothly varying signals defined on the nodes are of particular interest in various applications, and tend to be approximately bandlimited in the frequency basis. Sampling theory for graph signals deals with the problem of choosing the best subset of nodes for reconstructing a bandlimited signal from its samples. Most approaches to this problem require a computation of the frequency basis (i.e., the eigenvectors of the variation operator), followed by a search procedure using the basis elements. This can be impractical, in terms of storage and time complexity, for real datasets involving very large graphs. We circumvent this issue in our formulation by introducing quantities called graph spectral proxies, defined using the powers of the variation operator, in order to approximate the spectral content of graph signals. This allows us to formulate a direct sampling set selection approach that does not require the computation and storage of the basis elements. We show that our approach also provides stable reconstruction when the samples are noisy or when the original signal is only approximately bandlimited. Furthermore, the proposed approach is valid for any choice of the variation operator, thereby covering a wide range of graphs and applications. We demonstrate its effectiveness through various numerical experiments.
引用
收藏
页码:3775 / 3789
页数:15
相关论文
共 42 条
[1]  
Anis Aamir, 2014, 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), P3864, DOI 10.1109/ICASSP.2014.6854325
[2]  
[Anonymous], 2004, P ADV NEURAL INFORM
[3]  
[Anonymous], ARXIV150708822
[4]  
[Anonymous], 2014, Matrix analysis
[5]  
[Anonymous], 2009, CONVEX OPTIMIZATION
[6]  
[Anonymous], 2006, Advances in Neural Information Processing Systems
[7]   Emergence of scaling in random networks [J].
Barabási, AL ;
Albert, R .
SCIENCE, 1999, 286 (5439) :509-512
[8]   OPTIMAL PARTITIONS FOR EIGENVALUES [J].
Bourdin, Blaise ;
Bucur, Dorin ;
Oudet, Edouard .
SIAM JOURNAL ON SCIENTIFIC COMPUTING, 2009, 31 (06) :4100-4114
[9]  
Chao YH, 2015, 2015 PICTURE CODING SYMPOSIUM (PCS) WITH 2015 PACKET VIDEO WORKSHOP (PV), P60, DOI 10.1109/PCS.2015.7170047
[10]  
Chen SH, 2015, INT CONF ACOUST SPEE, P3392, DOI 10.1109/ICASSP.2015.7178600