Identifier

etd-042518-145218

Abstract

Accurate needle placement is critical to the success of needle-based interventions. Needle deflection due to tissue non-homogeneity and dynamic forces results in targeting error, potentially requiring repeated insertions. Real-time imaging enables closed-loop control of the needle during insertion, improving insertion accuracy. The needle localization algorithm proposed in this thesis models the needle as a parametric polynomial equation optimized to minimize beam bending energy relative to a set of observed needle coordinates. Simulated insertions using an MRI dataset show that the minimum bending energy model allows planning of subsequent imaging planes to capture the moving needle while estimating the shape of the needle with low error.

Publisher

Worcester Polytechnic Institute

Degree Name

MS

Department

Robotics Engineering

Project Type

Thesis

Date Accepted

2018-04-25

Accessibility

Unrestricted

Subjects

MRI, needle insertion, optimization

Share

COinS