Real-time intra-operative 3D tissue deformation recovery

Since the advent of laparoscopy, surgical technology has advanced on an exponential scale that has broadened the accessibility of the surgeon to the operative field with minimal incisions. Minimally Invasive Surgery (MIS) is carried out through natural body openings or small artificial incisions. It achieves its clinical goals with minimal inconvenience to patients, which results in reduced patient trauma, shortened hospitalisation, improved diagnostic accuracy and therapeutic outcome. With the introduction of robotic assisted MIS, the use of image guided surgical navigation is becoming increasingly popular, but it needs to handle non-rigid tissue deformation over the course of the procedure. In this paper, a probabilistic framework is presented that combines the strengths of different depth cues for tissue deformation recovery. The practicality of the technique is demonstrated using in vivo stereo laparoscopy data. Real-time intra-operative application of this technique has benefits for image based adaptive navigation and motion stabilisation in robotic assisted surgery.