Author

Ting-Hung Lin

Graduation Semester and Year

2008

Language

English

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Computer Science

Department

Computer Science and Engineering

First Advisor

Hua-Mei Chen

Abstract

Regridding was first introduced in viscous fluid registration for preventing folding of the transformation and for maintaining the admissible deformation field in large-deformation nonrigid image registration applications. We investigated the application of regridding to leading nonrigid image registration algorithms, including elastic, fluid, diffusion, curvature, and demons algorithms, and compared the performance and accuracy in each case.We also introduce a grid repairing mechanism based on the adaptive grid-generation method to prevent the transformation from folding. The grid repairing method can be used in conjunction with the proposed regridding scheme to set bounds on the Jacobian determinant of the transformation. We showed that our regridding and grid repairing method can outperform the original registration algorithms, particularly in large-deformation applications. In this dissertation, we also explain how the proposed method can improve the efficiencyability of the original registration algorithms for large-deformation applications and how the grid repairing method can be embedded in these algorithms.

Disciplines

Computer Sciences | Physical Sciences and Mathematics

Comments

Degree granted by The University of Texas at Arlington

Share

COinS