A Tree Based Recovery Algorithm for Block Sparse Signals
The structure of block sparsity in multi-band signals is prevalent. Performance of recovery algorithms that taking advantage of the block sparsity structure is promising in the compressed sensing framework. In this paper, the authors propose a binary tree based recovery algorithm for block-sparse signals, where they exploit the fact that each block may have zero and nonzero elements both. The proposed algorithm improves the current algorithms through iteratively separating the recovered blocks of signals into two smaller blocks. Therefore, greedy searching based algorithm is possible to obtain more accurate basis for signal recovery. Simulations are performed and the results show the improvements over current block-based recovery algorithms.