/Towards Vision Guided Retinal Vein Cannulation with an Actively Stabilized Handheld Robot

Towards Vision Guided Retinal Vein Cannulation with an Actively Stabilized Handheld Robot

Shohin Mukherjee
Master's Thesis, Tech. Report, CMU-RI-TR-17-20, Robotics Institute, Carnegie Mellon University, May, 2017

Download Publication (PDF)

Copyright notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author’s copyright. These works may not be reposted without the explicit permission of the copyright holder.

Abstract

In this thesis we describe work towards retinal vessel cannulation using an actively stabilized handheld robot, guided by monocular vision. Retinal vein cannulation is an incredibly delicate procedure because of the presence of physiological tremor in a surgeon’s hand which is typically greater than the diameter of the retinal vessels. Because of this reason, cannulation is not an approved medical procedure, even though it has the potential to treat several retinal diseases.

Our first contribution is an improved version of EySLAM, an eyeball motion tracking algorithm, that delivers 30 Hz real-time simultaneous localization and mapping of the human retina and vasculature during intraocular surgery. We implement graph based SLAM using the incremental matrix re-ordering and sparse nonlinear incremental graph optimization algorithm: ISAM2. This work also handles loop closures and demonstrates increased robustness to quick shaky motions and drift due to uncertainities in the motion estimation. We show considerable improvement over the earlier version of the algorithm.

Our second contribution is a monocular camera based surface reconstruction technique using laser beam scanning over the retina. We then use the reconstructed surface to estimate a coordinate transform between the 2D image plane coordinates and the the global 3D tracking frame of Micron. We introduce a motion scaling framework around the transformed targets for greater precision during approach. Experiments are conducted in a wet eye phantom to show the higher accuracy of the surface reconstruction as compared to standard stereo reconstruction. Further, experiments to show the increased surgical accuracy due to motion scaling are also carried out.

Our third contribution is to develop force control during cannulation with a new tool that integrates a micro-needle, a 2D force sensor and a laser probe. We then propose a integrated framework to carry out vessel cannulation with Micron by splitting the control mode into two stages: a motion scaling mode during approach to the vessel, and a force control mode after contact.

BibTeX Reference
@mastersthesis{Mukherjee-2017-22793,
author = {Shohin Mukherjee},
title = {Towards Vision Guided Retinal Vein Cannulation with an Actively Stabilized Handheld Robot},
year = {2017},
month = {May},
school = {Carnegie Mellon University},
address = {Pittsburgh, PA},
number = {CMU-RI-TR-17-20},
keywords = {Surgical robotics, Computer Vision, Micron, Retinal Vessel Cannulation, Surface Reconstruction, EyeSLAM},
}
2017-09-13T10:38:06+00:00