Toward Fluoroscopic Shape Reconstruction for Control of Steerable Medical Devices

Abstract

Control of surgical continuum robot manipulators and steerable needles requires accurate real-time sensing of tip position and/or shaft shape. Medical image feedback provides the most straightforward and widely available method of measuring device and clinical target positions and shapes during insertion or tissue manipulation. In this paper we present a method for automatic robot/needle curve segmentation from fluoroscopic images, as well as a method for 3D reconstruction of the curve using biplane fluoroscopy images.