Abstract
In this study, we proposed a new compressed sensing (CS) based image reconstruction method for photoacoustic tomography (PAT). To eliminate the finite aperture effect, the proposed method adopts the spatial impulse responses (SIRs) of the finite-sized flat transducer into the linear discrete PAT imaging model for CS. By using the nonlinear recovery algorithm based on convex optimization, PAT can be reconstructed with highly incomplete data. Therefore, the number of measurements and the system cost needed for a given image quality can be significantly reduced. In the mean time, retrospective restoration of the tangential resolution can be achieved because the SIR effect is incorporated in the CS. Simulation results demonstrate that this method not only reduces the data acquisition time but also improves the degraded tangential resolution for PAT with finite-sized flat transducers. © 2014 SPIE.