Statistical weighting of model-based optoacoustic reconstruction for minimizing artefacts caused by strong acoustic mismatch

A modified quantitative inversion algorithm is presented that minimizes the effects of internal acoustic reflections or scattering in tomographic optoacoustic images. The inversion procedure in our model-based algorithm consists in solving a linear system of equations in which each individual equation corresponds to a given position of the acoustic transducer and to a given time instant. Thus, the modification that we propose in this work consists in weighting each equation of the linear system with the probability that the measured wave is not distorted by reflection or scattering phenomena. We show that the probability that a reflected or scattered wave is detected at a given position and at a given instant is approximately proportional to the size of the area in which the original wave could have been generated, which is dependent on the position of the transducer and on the time instant, so that such probability can be used to weight each equation of the linear system. Thereby, the contribution of the waves that propagate directly to the transducer to the reconstructed images is emphasized. We experimentally test the proposed inversion algorithm with tissue-mimicking agar phantoms in which airgaps are included to cause reflections of the acoustic waves. The tomographic reconstructions obtained with the modification proposed herein show a clear reduction of the artefacts due to these acoustic phenomena with respect to the reconstructions yielded with the original algorithm. This performance is directly related to in-vivo small animal imaging applications involving imaging in the presence of bones, lungs, and other highly mismatched organs.