Photoacoustic image reconstruction using the pseudoinverse of the system matrix with the potential for real-time imaging

We discuss a method for using the pseudoinverse of the system matrix to perform photoacoustic image reconstruction. In our method, the regularization levels are set and the pseudoinverse matrices are calculated just once for all possible objects, so the reconstruction step consists only of a matrix-vector multiplication, which is very fast. We expect this method to work well in photoacoustic imaging because the dominant noise mechanism is usually transducer thermal noise, which is object independent. We find that this reconstruction method offers improvements in ideal observer signal-to-noise ratio, resolution, and the length of streak artifacts compared to standard filtered backprojection.