Hyperbolic harmonic brain surface registration with curvature-based landmark matching

摘要: Brain Cortical surface registration is required for inter-subject studies of functional and anatomical data. Harmonic mapping has been applied brain mapping, due to its existence, uniqueness, regularity numerical stability. In order improve the accuracy, sculcal landmarks are usually used as constraints registration. Unfortunately, constrained harmonic mappings may not be diffeomorphic produces invalid This work conquer this problem by changing Riemannian metric on target cortical a hyperbolic metric, so that guaranteed diffeomorphism while landmark enforced boundary matching condition. The computational algorithms based Ricci flow method heat diffusion. Experimental results demonstrate that, registrations always diffeomorphic, with higher qualities in terms alignment, curvature matching, area distortion overlapping region interests.