Introduction: The number of intravascular surgery performed in hospital is increasing every year. During the interventional procedure the cardiologist and neuro-surgeons are exposed to X-ray radiation [1]. In this work we aim to automate some standard steps performed during the interventional surgeries using a new robotics system developed at our institution.
Methods: The procedure starts with the selection of the required guide wire and catheter and the first step is to reach the Bovine Aortic Arch. The guide lines are developed for the robotic system to perform the procedure of advancing a guide catheter into the aortic arch, following an optimal path automatically, using image based catheter detection from X-ray images taken from a surgery trainer (VIST simulator – Mentice AB). The steps taken to carry out the procedure was demonstrated by experts and image data and motion was recorded. In our process X-ray imagery is captured by a camera and the images obtained are pre-processed [2] and the guide wire/catheter is recognized using a level by level thresholding technique. This data is fed to a Kalman filter to enhance the recognition rate.
Results: The pre-processing of the images from the simulated video feed is done with the concept of phase congruency [3] and the wire structures and the wire edge inside image were enhanced. Clustering and curve fitting operations were carried and the leading tip of the wire and remainder of the wire were identified. Using the Kalman filter the position of the wire was estimated in the successive frames. The various process parameters like states of the wire, actions [4], transition probability matrix and reward matrix were identified and using the Markov Decision Process the procedure was automated.
Conclusions: This research shows that the surgical tasks can be learned and automated using machine intelligence technique there by reducing exposure on surgeons.
Patient Care: Improve precision of surgical procedures using robotic device and machine intelligence.
Learning Objectives: Use of machine intelligence for improving image analysis of interventional surgery.
References: 1.Schneider, P.A., Endovascular Skills: Guidewires, Catheters, Balloon Angioplasty, Stents. 1998.
2.Honnorat, N., R. Vaillant, and N. Paragios, Robust guidewire segmentation through boosting, clustering and linear programming, in Biomedical Imaging: From Nano to Macro, 2010 IEEE International Symposium. 2010: Rotterdam. p. 924 - 927.
3.Kovesi, P. Phase Congruency Detects Corners and Edges. In The Australian Pattern Recognition Society Conference. 2003. Sydney.
4.Srimathveeravalli G, Li X, and Kesavadas T “Design and Fabrication of a Robotic Mechanism for Remote Steering and Positioning of Interventional Devices”, International Journal of Medical Robotics and Computer Aided Surgery, Vol. 6 (2), pp 160-170, June 2010.