作者: Kanwal K. Bhatia , Jo Hajnal , Alexander Hammers , Daniel Rueckert
DOI: 10.1007/978-3-540-75759-7_66
关键词:
摘要: The use of groupwise registration techniques for average atlas construction has been a growing area research in recent years. One particularly challenging component is finding scalable and effective similarity metrics; these do not always extend easily from pairwise metrics. This paper investigates possible choices metrics additionally proposes novel metric based on Normalised Mutual Information. described are quantitatively evaluated simulated 3D MR datasets, their performance compared to equivalent registration.