A multi-camera, multi-view system for training and skill assessment for robot-assisted surgery

Purpose: This paper introduces the concept of using an additional intracorporeal camera for the specific goal of training and skill assessment and explores the benefits of such an approach. This additional camera can provide an additional view of the surgical scene, and we hypothesize that this additional view would improve surgical training and skill assessment in robot-assisted surgery. Methods: We developed a multi-camera, multi-view system, and we conducted two user studies (N= 12) to evaluate its effectiveness for training and skill assessment. In the training user study, subjects were divided into two groups: a single-view group and a dual-view group. The skill assessment study was a within-subject study, in which every subject was shown single- and dual view recorded videos of a surgical training task, and the goal was to count the number of errors committed in each video. Results: The results show the effectiveness of using an additional intracorporeal camera view for training and skill assessment. The benefits of this view are modest for skill assessment as it improves the assessment accuracy by approximately 9%. For training, the additional camera view is clearly more effective. Indeed, the dual-view group is 57% more accurate than the single-view group in a retention test. In addition, the dual-view group is 35% more accurate and 25% faster than the single-view group in a transfer test. Conclusion: A multi-camera, multi-view system has the potential to significantly improve training and moderately improve skill assessment in robot-assisted surgery. One application of our work is to include an additional camera view in existing virtual reality surgical training simulators to realize its benefits in training. The views from the additional intracorporeal camera can also be used to improve on existing surgical skill assessment criteria used in training systems for robot-assisted surgery.