The application of dictionary based compressed sensing for photoacoustic image

Abstract

Restrictions of the hardware conditions and spatial size usually limit the number of the measurements in photo acoustic imaging which will finally degrade the quality of the reconstructed image with the back projection algorithm. In order to recover larger number of measurements from incomplete ones, a compressed sensing (CS) based method was proposed. Different from most existed CS-based photoacoustic reconstruction method, the transform matrix for converting the measurement data to their compressed version is obtained by learning a dictionary with the K-SVD method. Visual assessment and quantitative evaluations in terms of the mean squared error (MSE) and the peak signal-to-noise ratio (PSNR) demonstrate the superiorities of our proposed method.