An efficient algorithm for total variation sparse-view photoacoustic image reconstruction

Abstract:
Photoacoustic image reconstruction from sample data is a very hot research area. Total variation regularization is popular in image reconstruction due to its edge-preserving property. In this paper, we use a total variation regularization with alternating direction method for sparse photoacoustic image reconstruction. The numerical results show that this algorithm performs favorably in comparison to other methods while a low number of transducers was. Quantitative comparison demonstrates that, at the same condition, the used algorithm leads to a peak-signal-to-noise ratio for about two times of other methods, the relative errors of TVAL3 are almost one-tenth of L1 norm based methods and it keeps the lowest computation time.