Photoacoustic Imaging Reconstruction Algorithm Based on a Combined First and Second Order Total Variation Regularization

In photoacoustic tomography (PAT), total variation (TV) based iteration algorithm is reported to have a good performance in PAT image reconstruction. However, classical TV based algorithm fails to preserve the edges and texture details of the image especially for sparse-view scanning. To improve the reconstruction quality of total variation (TV) based two-dimensional photoacoustic imaging (PAI) reconstruction methods, we incorporate the high order TV penalty term into PAT iterative algorithm. This method involves the combination of the first and second order of TV (FSTV), which can effectively eliminate the painting like artifacts of TV based methods. The effective Barzilai-Borwein step size selection method is used to solve the optimization problem so that the calculation and convergence of the algorithm are greatly accelerated. Numerical results demonstrate the proposed PAI reconstruction method is superior to TV-based method in both visual quality and quantitative indicators such as PSNR, convergence speed and robustness. In-vitro experiments results show that the FSTV yielded improved image quality over the TV in the cases studied in this manuscript.