A brief talk on compression perception (27): Sparse Adaptive matching tracking for compression-aware reconstruction algorithms (SAMP)

Main content:

1. Algorithmic flow of Samp
2. Samp's MATLAB implementation
3. Experiment and result of one-dimensional signal
4. Experiment and result of probability relationship between sparsity K and reconstruction success

First, the SAMP algorithm flow

As mentioned above, most of the OMP and its pre-modification algorithms need to know the sparsity of the signal k, and in practice this is generally unknown, based on this background, sparse adaptive matching tracking (sparsity Adaptive MP) was proposed. Samp does not need to know the sparsity K, in the iterative cycle, according to the new residuals and the old residual difference to determine the number of selected atoms.

Samp Algorithm Flow:

Second, the SAMP of MATLAB implementation (CS_SAMP.M)

?

Experiment and result of three or one-D signal

?

Experiment and results of the relationship between sparse number k and the probability of reconstruction success

Vi. Articles of Reference

http://blog.csdn.net/jbb0523/article/details/45675735

A brief talk on compression perception (27): Sparse Adaptive matching tracking for compression-aware reconstruction algorithms (SAMP)