System and Method for Corpus Callosum Segmentation in Magnetic Resonance Images

摘要

A method and system for segmentation of the corpus callosum in MR brain images is disclosed. The method utilizes an active shape model (ASM) with confidence weighting to iteratively adjust an initial corpus callosum contour to define a boundary of the corpus callosum in an MR image. An ASM is used to determine a displacement value in a perpendicular direction of the corpus callosum contour for each node of the corpus callosum contour. The displacement value for each node is then weighted based on a confidence of that displacement value. The ASM is then fitted to the adjusted contour. These steps are iteratively performed until the contour converges to define the corpus callosum boundary. This boundary can be further refined based on intensity distributions in object and background regions defined by the boundary.