Pose Estimation of an Ablation Catheter

Abstract :

Surgical catheter robots are used for the cardiovascular diagnostics or treatments. For modelling a catheteraorta interaction, joint probability densities are used, which are represented by mixture of Gaussians. During training of the joint probability densities, different features of the catheter, e.g., the catheter tip position, catheter shape, catheter bending angle are captured as the input data to the model. In order to extract those features of the catheter in a 3D aorta mock up, the pose estimation is implemented. From the two HD cameras, the pixels of the catheter distal section are detected. Then based on the position of the catheter tip, an interactive graph cut method is used for extracting the catheter center points, which are fitted by the B-Spline curves. According to the fitted B-splines, the 3D shape of the catheter is estimated and then presented in a simulated aorta mesh.