randomwalk94 / tomp Goto Github PK

The scrips in this repository are numerical implementations of the paper (TOMP_Siam.pdf)

Thresholding Orthogonal Matching Pursuit for Imaging Applications, Hai Le, Alexei Novikov.

The main.m file (the only file to run) compares three algorithms: SquareRoot_LASSO, Stagewise Orthogonal Matching Pursuit (StOMP), and Thresholding Orthogonal Matching Pursuit (TOMP) for the sparse recovery problem:

Ax=y

where A is a measurement matrix, y is a data vector, and x is the sparse vector that we want to recover.

The settings for the system Ax = y come from the settings of Passive Array Imaging from the paper https://arxiv.org/abs/1908.01479 where A is a highly coherent imaging matrix (nearby normalized column vectors are almost parallel to each other).