Disease Classification with Hippocampal Shape Invariants

Boris Gutman, Yalin Wang, Jonathan Morra, Arthur W. Toga and Paul M. Thompson

Abstract

We present the first Support Vector Machine classification study using the feature space of shape invariants of hippocampal sur- faces. Our shape invariants are based on rotationally invariant proper- ties of spherical harmonics (SPH). A global conformal map is used for parameterization. Leave-one-out testing on 49 Alzheimer(AD) and 63 elderly control subjects yielded 75.5% sensitivity and 87.3% specificity with 82.1% correct overall.

Figures (click on each for a larger version):

Related Publications

B. Gutman, Y. Wang, J. Morra, P.M. Thompson, A.W. Toga, Disease Classification with Hippocampal Shape Invariants, MICCAI 2008 Workshop on the Computational Anatomy and Physiology of the Hippocampus, pp. 76-86
B. Gutman, Y. Wang, J. Morra, P.M. Thompson, A.W. Toga, Disease Classification with Hippocampal Shape Invariants, Hippocampus 19, 2009, pp. 572-578