Intrinsic shape context descriptors for deformable shapes

Iasonas Kokkinos (1, 2) , Michael M. Bronstein (3) , Roee Litman (4) , Alexander M. Bronstein (4)

1 GALEN - Organ Modeling through Extraction, Representation and Understanding of Medical Image Content
2 CVN - Centre de vision numérique
3 Faculty of Informatics [Lugano]
4 Department of Electrical Engineering

In this work, we present intrinsic shape context (ISC) descriptors for 3D shapes. We generalize to surfaces the polar sampling of the image domain used in shape contexts: for this purpose, we chart the surface by shooting geodesic outwards from the point being analyzed; `angle' is treated as tantamount to geodesic shooting direction, and radius as geodesic distance. To deal with orientation ambiguity, we exploit properties of the Fourier transform. Our charting method is intrinsic, i.e., invariant to isometric shape transformations. The resulting descriptor is a meta-descriptor that can be applied to any photometric or geometric property field defined on the shape, in particular, we can leverage recent developments in intrinsic shape analysis and construct ISC based on state-of-the-art dense shape descriptors such as heat kernel signatures. Our experiments demonstrate a notable improvement in shape matching on standard benchmarks.

Domaines: Vision par ordinateur et reconnaissance de formes [cs.CV]

Iasonas Kokkinos, Michael M. Bronstein, Roee Litman, Alexander M. Bronstein. Intrinsic shape context descriptors for deformable shapes. CVPR 2012 - IEEE Conference on Computer Vision and Pattern Recognition, Jun 2012, Providence, United States. pp.159-166, ⟨10.1109/CVPR.2012.6247671⟩. ⟨hal-00857572⟩