Dataset supporting: Distribution-informed and wavelength-flexible data-driven photoacoustic oximetry

This data set comprises simulated photoacoustic imaging data of various synthetic geometries (training_processed) and experimentally acquired data of two blood flow phantoms, of mice, and of human forearms (test_final). A complete description of these datasets is provided in the associated paper, which is available open access and on the preprint server arXiv. Furthermore, the data contain pre-trained models on each of the training datasets with various sparsity levels (models_LSTM) and a few examples of the raw simulation outputs produced by the SIMPA toolkit before preprocessing (sample_images). Unfortunately, it is not feasible to share all raw data due to its file size (>1.5 TB), but all results presented in this work are reproducible from the provided processed data.