Abstract | ||
---|---|---|
Atlas-based brain image segmentation quality is difficult to assess in the absence of reference target segmentations. We propose a measure of segmentation success based on transforming the atlas label twice: once by registering the atlas to the target and a second time by registering the target to the atlas. Each registration represents transformations by free-form deformations. The overlap between the twice-transformed label and the original ('mirror overlap') correlates with the forward overlap (between the once-transformed label and a target reference), especially for subcortical structures. Using mirror overlap as an atlas selection criterion results in improved segmentations versus random selection. |
Year | DOI | Venue |
---|---|---|
2009 | 10.1109/ISBI.2009.5193272 | ISBI |
Keywords | Field | DocType |
free-form deformation,atlas-based brain image segmentation,random selection,improved segmentation,mirror method,atlas selection criterion result,segmentation quality,atlas label propagation,once-transformed label,reference target segmentation,twice-transformed label,target reference,segmentation success,image registration,data mining,atlas,image segmentation,biomedical imaging,brain imaging,magnetic resonance image,magnetic resonance imaging,correlation | Computer vision,Pattern recognition,Label propagation,Computer science,Segmentation,Medical imaging,Free-form deformation,Image segmentation,Atlas (anatomy),Selection criterion,Artificial intelligence,Image registration | Conference |
Citations | PageRank | References |
1 | 0.35 | 6 |
Authors | ||
5 |
Name | Order | Citations | PageRank |
---|---|---|---|
R A Heckemann | 1 | 292 | 10.47 |
A Hammers | 2 | 399 | 18.98 |
P Aljabar | 3 | 393 | 19.19 |
Daniel Rueckert | 4 | 9338 | 637.58 |
J. V. Hajnal | 5 | 359 | 17.44 |