Part-based probabilistic point matching using equivalence constraints

Graham McNeill, S. Vijayakumar

Research output: Chapter in Book/Report/Conference proceedingConference contribution


Correspondence algorithms typically struggle with shapes that display part-based
variation. We present a probabilistic approach that matches shapes using independent part transformations, where the parts themselves are learnt during matching. Ideas from semi-supervised learning are used to bias the algorithm towards finding ‘perceptually valid’ part structures. Shapes are represented by unlabeled point sets of arbitrary size and a background component is used to handle occlusion, local dissimilarity and clutter. Thus, unlike many shape matching techniques, our approach can be applied to shapes extracted from real images. Model parameters are estimated using an EM algorithm that alternates between finding a soft correspondence and computing the optimal part transformations using Procrustes analysis.
Original languageEnglish
Title of host publicationProc. Advances in Neural Information Processing Systems (NIPS '06), Vancouver, Canada
Publication statusPublished - 2006


Dive into the research topics of 'Part-based probabilistic point matching using equivalence constraints'. Together they form a unique fingerprint.

Cite this