Multiscale MRF optimization for robust registration of 2D biological data

摘要: Discrete formulations of image registration offer the promise dense deformations via optimizations robust to large motions or poor initialization. However, many available efficient algorithms are not well suited medical biological data. We propose a novel multiscale Markov Random Field formulation for registration, which reduces number labels needed at each scale while preserving ability represent dense, fine-grained feature matches. The nature algorithm also allows arbitrary sub-voxel accuracy, and we further simple extension grants measure rotational invariance an matching term.