Goal-conditioned reinforcement learning for ultrasound navigation guidance

arXiv:2405.01409v1 Announce Type: new
Abstract: Transesophageal echocardiography (TEE) plays a pivotal role in cardiology for diagnostic and interventional procedures. However, using it effectively requires extensive training due to the intricate nature of image acquisition and interpretation. To enhance the efficiency of novice sonographers and reduce variability in scan acquisitions, we propose a novel ultrasound (US) navigation assistance method based on contrastive learning as goal-conditioned reinforcement learning (GCRL). We augment the previous framework using a novel contrastive patient batching method (CPB) …

abstract acquisition acquisitions arxiv cs.ai cs.cv diagnostic efficiency guidance however image interpretation nature navigation novel pivotal reduce reinforcement reinforcement learning role training type