Sparse representation of piecewise stationary signals based on overcomplete dictionaries

作者：Wang Tianjiang^*; Chang Baoxian; Yang Zhen; Zheng Baoyu

来源：Journal of Computational Information Systems, 2014, 10(11): 4637-4648.

DOI：10.12733/jcis10257

• Autocorrelation functions
• Discrete Cosine Transform(DCT)
• Nonlinear approximation
• Orthogonal matching pursuit
• Over-complete dictionaries
• Piecewise stationaries
• Signal representations
• Sparse representation

In recent years there has been growing interests in sparse representation of signals based on overcomplete dictionaries. Recent activities in this field mainly concentrate on the study of dictionary design algorithm and sparse decomposition algorithm. In this paper, we propose a novel sparse representation of piecewise stationary signals based on overcomplete dictionaries. An effective overcomplete dictionary is constructed by the autocorrelation function model of piecewise stationary process. Furthermore, a sparse decomposition algorithm in accordance with nonlinear approximation is designed to obtain sparse representation of piecewise stationary signals, which has lower computational cost than orthogonal matching pursuit and basis pursuit. Benefiting from algebraic structure of atoms, the proposed method has higher sparsity of signal representation and better reconstruction performance than sparse representation of piecewise stationary signals based on overcomplete Discrete Cosine Transform (DCT) dictionaries. Experimental results show its stability and applicability, so it can preferably apply to compressed sensing.

• 出版日期2014
• 单位南京邮电大学