Fast iterative reconstruction algorithm for microwave-induced thermoacoustic imaging

In this paper we propose a fast iterative reconstruction algorithm for microwave-induced thermoacoustic (TA) imaging. Instead of calculating the forward and back projections in real time, we construct a library of the impulse responses of individual pixels in advance. During reconstruction, only covariance calculations are performed to reconstruct the intensity of each pixel. In addition, we propose to use a Wiener filter based-deconvolution to accelerate convergence, enabling even single iteration construction in some scenarios. We validate the proposed method both with simulations as well as experiments. Compared to the time reversal method, the proposed fast iterative reconstruction demonstrates better image quality with reduced computational time.