Quickly matching pursuit algorithm based atom clusters
原子簇快速匹配追踪算法

This paper proposed a new quickly matching pursuit algorithm based on atomic correlation for sparse representation. Firstly, it clustered the atoms of redundant dictionary into M multi-atoms sets after making full use of the atomic correlation in the iterative process. Then at each iteration, it selected one multi-atoms set as the target atom cluster. Finally, it got some atoms that met the matching conditions from the target atom cluster and the signal approximation. Experimental results for 1D sparse signal show that the calculation speed of the algorithm increases significantly compared with MP's. Meanwhile, the approximation performances of the proposed method are comparable with those of the traditional matching pursuit methods.