Projection Onto Convex Sets (POCS) method for photoacoustic tomography with a non negative constraint

Abstract

Photoacoustic imaging is a biomedical imaging modality capable of early cancer detection. In this paper, we proposed a novel iterative Projections Onto Convex Sets (POCS) method for improving photoacoustic reconstruction. This method aims to obtain a non negative pressure distribution satisfying the measured signals. This POCS method is performed in the Fourier Bessel space avoiding matrix inversions in the projections, speeding up projections and is capable of handling the large data sets present in photoacoustic imaging. The numerical experiments performed showed that improved reconstruction was obtained with a few iterations together with the recovery of some lost information.