TY - GEN
T1 - Hyperbolic harmonic brain surface registration with curvature-based landmark matching
AU - Shi, Rui
AU - Zeng, Wei
AU - Su, Zhengyu
AU - Wang, Yalin
AU - Damasio, Hanna
AU - Lu, Zhonglin
AU - Yau, Shing Tung
AU - Gu, Xianfeng
PY - 2013
Y1 - 2013
N2 - Brain Cortical surface registration is required for inter-subject studies of functional and anatomical data. Harmonic mapping has been applied for brain mapping, due to its existence, uniqueness, regularity and numerical stability. In order to improve the registration accuracy, sculcal landmarks are usually used as constraints for brain registration. Unfortunately, constrained harmonic mappings may not be diffeomorphic and produces invalid registration. This work conquer this problem by changing the Riemannian metric on the target cortical surface to a hyperbolic metric, so that the harmonic mapping is guaranteed to be a diffeomorphism while the landmark constraints are enforced as boundary matching condition. The computational algorithms are based on the Ricci flow method and hyperbolic heat diffusion. Experimental results demonstrate that, by changing the Riemannian metric, the registrations are always diffeomorphic, with higher qualities in terms of landmark alignment, curvature matching, area distortion and overlapping of region of interests.
AB - Brain Cortical surface registration is required for inter-subject studies of functional and anatomical data. Harmonic mapping has been applied for brain mapping, due to its existence, uniqueness, regularity and numerical stability. In order to improve the registration accuracy, sculcal landmarks are usually used as constraints for brain registration. Unfortunately, constrained harmonic mappings may not be diffeomorphic and produces invalid registration. This work conquer this problem by changing the Riemannian metric on the target cortical surface to a hyperbolic metric, so that the harmonic mapping is guaranteed to be a diffeomorphism while the landmark constraints are enforced as boundary matching condition. The computational algorithms are based on the Ricci flow method and hyperbolic heat diffusion. Experimental results demonstrate that, by changing the Riemannian metric, the registrations are always diffeomorphic, with higher qualities in terms of landmark alignment, curvature matching, area distortion and overlapping of region of interests.
UR - https://www.scopus.com/pages/publications/84901259437
U2 - 10.1007/978-3-642-38868-2_14
DO - 10.1007/978-3-642-38868-2_14
M3 - Conference contribution
C2 - 24683966
AN - SCOPUS:84901259437
SN - 9783642388675
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 159
EP - 170
BT - Information Processing in Medical Imaging - 23rd International Conference, IPMI 2013, Proceedings
T2 - 23rd International Conference on Information Processing in Medical Imaging, IPMI 2013
Y2 - 28 June 2013 through 3 July 2013
ER -