Signal domain adaptation network for limited-view optoacoustic tomography

Anna Klimovskaia Susmelj, Berkan Lafci, Firat Ozdemir, Neda Davoudi, Xosé-Luís Dean-Ben, Fernando Pérez-Cruz, Daniel Razansky

Abstract: Highlights•Signal domain adaptation network is an unsupervised approach to remove limited-view artifacts in optoacoustic imaging via a combination of domain adaptation and sides prediction networks.•Domain adaptation network reduces the domain gap between simulations and experiments, allowing to improve the quality of the prediction model trained on simulated and applied to experimental data.•Sides prediction networks impute the signals corresponding to the missing part of the tomographic geometry.•The detected along with the predicted (missing) signals are subsequently used to reconstruct optoacoustic images.•Four metrics are used for assessing the performance of the proposed methods in reducing image artifacts resulting from limited tomographic coverage.