Journal
PHYSICS IN MEDICINE AND BIOLOGY
Volume 52, Issue 16, Pages 4721-4738Publisher
IOP PUBLISHING LTD
DOI: 10.1088/0031-9155/52/16/001
Keywords
-
Funding
- NCI NIH HHS [R01 CA 98524-01] Funding Source: Medline
Ask authors/readers for more resources
This paper presents a new concept to automatically detect the neighborhood in an image where deformable registration is mis-performing. Specifically, the displacement vector field (DVF) from a deformable image registration is substituted into a finite-element-based elastic framework to calculate unbalanced energy in each element. The value of the derived energy indicates the quality of the DVF in its neighborhood. The new voxel-based evaluation approach is compared with three other validation criteria: landmark measurement, a finite element approach and visual comparison, for deformable registrations performed with the optical-flow-based 'demons' algorithm as well as thin-plate spline interpolation. This analysis was performed on three pairs of prostate CT images. The results of the analysis show that the four criteria give mutually comparable quantitative assessments on the six registration instances. As an objective concept, the unbalanced energy presents no requirement on boundary constraints in its calculation, different from traditional mechanical modeling. This method is automatic, and at voxel level suitable to evaluate deformable registration in a clinical setting.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available