Ultrasound Image simulation

Recently, Cardiovascular Imaging and Dynamics lab proposed a methodology toward ultra-realistic simulation of cardiac ultrasound data with important applications in machine learning as an infinite source of input data to train models. Despite the realism of these synthetic data sets, an important shortcoming is the lack of realistic artefacts that are common in real life and thus instrumental towards proper training of models. Within this PhD work, methodologies will be developed to incorporate these artefacts based on physical models of wave propagation in soft tissue.