Date of Award
2008
Degree Type
Thesis
Degree Name
Master of Engineering Science
Program
Electrical and Computer Engineering
Supervisor
Dr. Rajnikant V. Patel
Abstract
Three-dimensional (3D) ultrasound-guided needle insertion has recently received much attention in minimally invasive percutaneous procedures specifically in needle steering techniques. 3D ultrasound provides limited spatial perception due to the distorted appearance of tissue and needles. The low quality imaging when combined with the curvature in the needles makes accurate targeting difficult for clinicians or surgical robots. In this thesis a novel technique for real-time segmentation and tracking of curved needles in 3D ultrasound images is presented. A minimally invasive test-bed which includes software and hardware components to capture and reconstruct 3D volumetric images from 2D cross-sectional laparoscopic ultrasound images and to robotically insert needles into tissue is developed. A new parameterization for the generalized Radon transform is introduced and implemented on a graphics processing unit using NVIDIA’s Compute Unified Device Architecture (CUDA) programming. Apart from the main segmentation algorithm an off-line electromagnetic based tracking system is also developed for comparison of the image-based segmentation results. The effectiveness of the algorithm is demonstrated on ex-vivo animal tissue 3D-US images of different needles carried by a laparoscopic needle holder mounted on a minimally invasive surgical robotic test-bed.
Recommended Citation
Sadeghi Neshat, Hamid Reza, "Real-Time Segmentation of Curved Needles For 3D Ultrasound Image Guidance" (2008). Digitized Theses. 4738.
https://ir.lib.uwo.ca/digitizedtheses/4738