Image quality assessment for machine learning tasks using meta-reinforcement learning

Medical imaging is used extensively for diagnostic and therapeutic procedures in medicine, whether they be interventional or non-interventional in nature. Several such diagnostic, navigational or therapeutic tasks in the clinical workflow rely on medical images where they inform the clinician's judgement, directly or via derived measurements. Medical imaging is increasingly being used as a navigational aid to guide surgical and other interventional procedures, such as for prostate biopsies (brown_prostate_biopsy), liver resections (simpson_liver_resection), and brain resections (kondziolka_brain_resection). Treatment planning, for example radiotherapy planning, relies heavily on pre-operative medical images (dirix_mr_radiotherapy; liney_mr_radiotherapy_q). Moreover, imaging is commonly used for diagnostic clinical tasks whether the task is performed manually by humans or automated using computer aided diagnosis.