GPU-accelerated, gradient-free MI deformable registration for atlas-based MR brain image segmentation

• Published in: 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops pp. 141 - 148
• Main Authors: Xiao Han, Hibbard, Lyndon S, Willcut, Virgil
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.06.2009
• Subjects: Acceleration; Application software; Brain; Collision mitigation; Deformable models; Image registration; Image segmentation; Magnetic resonance; Neuroscience; Optimization methods
• ISBN: 1424439949; 9781424439942
• ISSN: 2160-7508
• DOI: 10.1109/CVPRW.2009.5204043

Brain structure segmentation is an important task in many neuroscience and clinical applications. In this paper, we introduce a novel MI-based dense deformable registration method and apply it to the automatic segmentation of detailed brain structures. Together with a multiple atlas fusion strategy, very accurate segmentation results were obtained, as compared with other reported methods in the literature. To make multi-atlas segmentation computationally feasible, we also propose to take advantage of the recent advancements in GPU technology and introduce a GPU-based implementation of the proposed registration method. With GPU acceleration it takes less than 8 minutes to compile a multi-atlas segmentation for each subject even with as many as 17 atlases, which demonstrates that the use of GPUs can greatly facilitate the application of such atlas-based segmentation methods in practice.