Multi-Object Analysis of Volume, Pose, and Shape Using Statistical Discrimination

Abstract— One goal of statistical shape analysis is the discrimination between two populations of objects. In this paper, we present results of discriminant analysis on multi-object sets, a problem not yet sufficiently explored. Our choice of discriminant method is the distance weighted discriminant (DWD) because of its generalization ability in high dimensional, low sample size settings. Using an unbiased, soft discrimination score we can associate a statistical hypothesis test with the discrimination results. Furthermore, via the average discriminating axis, we can visualize the differences between the populations. We explored the effectiveness of different choices of features as input to the discriminant analysis. In a multi-object setting, it may not be solely shape that is significant in discriminating between populations. The relative pose of objects to each other may also be of interest, which implies that the alignment procedure preceding the shape analysis must be selected...

• Type: Journal
• Year: 2010
• Where: PAMI
• Authors: Kevin Gorczowski, Martin Styner, Ja-Yeon Jeong, J. S. Marron, Joseph Piven, Heather Cody Hazlett, Stephen M. Pizer, Guido Gerig
• Keywords: Discriminant Analysis, PAMI 2010, Shape Analysis, Statistical Shape Analysis