TY - GEN
T1 - Block-based statistics for robust non-parametric morphometry
AU - Chen, Geng
AU - Zhang, Pei
AU - Li, Ke
AU - Wee, Chong Yaw
AU - Wu, Yafeng
AU - Shen, Dinggang
AU - Yap, Pew Thian
N1 - Publisher Copyright:
© Springer International Publishing Switzerland 2015.
PY - 2015
Y1 - 2015
N2 - Automated algorithms designed for comparison of medical images are generally dependent on a sufficiently large dataset and highly accurate registration as they implicitly assume that the comparison is being made across a set of images with locally matching structures. However, very often sample size is limited and registration methods are not perfect and may be prone to errors due to noise, artifacts, and complex variations of brain topology. In this paper, we propose a novel statistical group comparison algorithm, called block-based statistics (BBS), which reformulates the conventional comparison framework from a non-local means perspective in order to learn what the statistics would have been, given perfect correspondence. Through this formulation, BBS (1) explicitly considers image registration errors to reduce reliance on high-quality registrations, (2) increases the number of samples for statistical estimation by collapsing measurements from similar signal distributions, and (3) diminishes the need for large image sets. BBS is based on permutation test and hence no assumption, such as Gaussianity, is imposed on the distribution. Experimental results indicate that BBS yields markedly improved lesion detection accuracy especially with limited sample size, is more robust to sample imbalance, and converges faster to results expected for large sample size.
AB - Automated algorithms designed for comparison of medical images are generally dependent on a sufficiently large dataset and highly accurate registration as they implicitly assume that the comparison is being made across a set of images with locally matching structures. However, very often sample size is limited and registration methods are not perfect and may be prone to errors due to noise, artifacts, and complex variations of brain topology. In this paper, we propose a novel statistical group comparison algorithm, called block-based statistics (BBS), which reformulates the conventional comparison framework from a non-local means perspective in order to learn what the statistics would have been, given perfect correspondence. Through this formulation, BBS (1) explicitly considers image registration errors to reduce reliance on high-quality registrations, (2) increases the number of samples for statistical estimation by collapsing measurements from similar signal distributions, and (3) diminishes the need for large image sets. BBS is based on permutation test and hence no assumption, such as Gaussianity, is imposed on the distribution. Experimental results indicate that BBS yields markedly improved lesion detection accuracy especially with limited sample size, is more robust to sample imbalance, and converges faster to results expected for large sample size.
UR - http://www.scopus.com/inward/record.url?scp=84955287169&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-28194-0_8
DO - 10.1007/978-3-319-28194-0_8
M3 - 会议稿件
AN - SCOPUS:84955287169
SN - 9783319281933
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 62
EP - 70
BT - Patch-Based Techniques in Medical Imaging - First st International Workshop, Patch-MI 2015 Held in Conjunction with MICCAI 2015, Revised Selected Papers
A2 - Coupé, Pierrick
A2 - Munsell, Brent
A2 - Wu, Guorong
A2 - Zhan, Yiqiang
A2 - Rueckert, Daniel
PB - Springer Verlag
T2 - 1st International Workshop on Patch-Based Techniques in Medical Imaging, Patch-MI 2015
Y2 - 9 October 2015 through 9 October 2015
ER -