Improving photoacoustic imaging through the skull using deep learning: Effect of shear waves

Human brain mapping has become one of the most exciting fields of research for medical imaging. Among other imaging techniques, photoacoustic computed tomography (PACT) has shown great promise due to its rich optical absorption contrast, high spatial and temporal resolutions, and relatively deep penetration. However, before the acoustic signal makes it to the transducer array for image reconstruction, it becomes distorted by the porous skull, resulting in blurry images. In recent years, deep learning (DL) has been proposed as a solution to achieve images of higher quality, with U-Net being a common architecture. In this numerical study, two types of images were generated using the k-Wave toolbox in MATLAB: one that only considered longitudinal waves, and another that added the effects of shear waves to reflect more realistic conditions of the skull. Afterwards, our U-Net model was trained on the longitudinal wave dataset and tested on the longitudinal+shear wave dataset. The results still show improved images, although the loss metrics have slightly degraded. These findings demonstrate U-Net’s robustness and provide insights into future architectural enhancements.