Directed Feedback for Simulation-based Training in Robot-assisted Minimally Invasive Surgery
Surgical educators have recommended the need for objective and automated teaching tools for technical skills training. Current training platforms provide automated or manual global assessments that are based on postoperative performance review. Approaches like these lack task segment specific feedback that informs the trainee of where, why and what went wrong that led to the lower performance scores, if at all. Coaching with targeted feedback is crucial not only for new trainees, but also equally critical for experienced surgeons to retain their skills and maintain higher levels of performance. The objective of my thesis project is to develop a virtual coach that can provide directed feedback to the trainee. This includes step-by-step performance evaluation, skill deficiency or error detection, and relevant demonstration presentation. Our framework mainly consists of three components – a) task segmentation, b) segment-level evaluation, and c) feedback generation for targeted practice and improvement.