Recorded demonstrations of robot-assisted minimally invasive surgery (RMIS) have been used for surgical skill assessment [7], development of finite state machines for automation [13, 25], learning from demonstration (LfD) [29], and calibration [22]. Intuitive Surgical's da Vinci robot facilitated over 570, 000 procedures in 2014 [11]. There are proposals to record all of Intuitive's RMIS procedures similar to flight data recorders ("black boxes") in airplanes [12], which could lead to a proliferation of data. While these large datasets have the potential to facilitate learning and autonomy; the length and variability of surgical trajectories pose a unique challenge. Each surgical trajectory may represent minutes of multi-modal observations, may contain loops (failures and repetitions until achieving the desired result), and even identical procedures can vary due to differences in the environment. In this setting, typical techniques for establishing spatial and temporal correspondence that employ continuous deformations can be unreliable (e.g., Dynamic Time Warping [14] and spline-based registration [31]).
展开▼
