Fast Thermoacoustic Imaging Based on Compressive Sensing Applying an Effective Algorithm

Abstract

This work reports the application of a new algorithm called FISTA (fast iterative shrinkage-thresholding algorithm) to solve 3D thermoacoustic imaging problem based on compressive sensing. It shows much higher convergence rate than GPSR (gradient projection for sparse reconstruction) algorithm which is normally used for solving thermoacoustic imaging problem based on compressive sensing. A simulation work is done to prove the fact that FISTA is of faster convergence rate than GPSR. To achieve a certain recovered image quality, FISTA needs much less iteration steps than GPSR algorithm.