PhD defense : 3D pose estimation for the control of flexible instruments in robotic endoscopic surgery

PhD defense : Paolo CABRAS

Team : AVR

Title : 3D pose estimation for the control of flexible instruments in robotic endoscopic surgery

Abstract : Thanks to their dexterity and compliance, flexible systems can reach distal body zone in a real non-invasive way through natural orifice allowing no-scar or single scar surgery. Being able to measure the 3D position of such instruments can be useful for various tasks, such as controlling automatically the robotized instruments or providing gesture guidance. In this thesis, we propose two automatic methods to infer the 3D pose of a single bending section instrument using only the images provided by the monocular camera embedded at the tip of the endoscope. Both methods relies on colored markers attached onto the bending section. The image of the instrument is segmented using a graph-based method and the corners of the markers are extracted by detecting the color transition along Bézier curves fitted on edge points. These features are then used to estimate the 3D pose of the instrument using an adaptive model that takes into account backlash between the instrument and its housing channel. Strong model uncertainties can affect the result of such model-based method. Therefore, two learning approaches have been studied that approximate image features-to-3D function according to a training set: Radial Basis Function (RBF) Network and Locally Weighted Regression (LWR). The first proposal approximate the function as a weighted sum of Gaussians whose weights are computed according to a global criteria. The other approach, instead, performs a linear approximation according to local information of the training set. The proposed methods are validated on a robotic experimental cell and in in-vivo sequences. Advantages and inconveniences of each method are presented and commented.

This thesis was supervised by Christophe Doignon (professor at the University of Strasbourg).

The presentation will take place on Wednesday, February 24th starting at 1.30pm in the Lindbergh lecture hall at IRCAD. This thesis will be supported in english.