Context-dependent kernel design for object matching and recognition

Hichem Sahbi (1) , Jean-Yves Audibert (2, 3, 4) , Jaonary Rabarisoa (2, 3, 4) , Renaud Keriven (2, 3, 4)

1 LTCI - Laboratoire Traitement et Communication de l'Information
2 CERTIS - Centre d'Enseignement et de Recherche en Technologies de l'Information et Systèmes
3 imagine [Marne-la-Vallée]
4 LIGM - Laboratoire d'Informatique Gaspard-Monge

The success of kernel methods including support vector networks (SVMs) strongly depends on the design of appropriate kernels. While initially kernels were designed in order to handle fixed-length data, their extension to unordered, variable-length data became more than necessary for real pattern recognition problems such as object recognition and bioinformatics. We focus in this paper on object recognition using a new type of kernel referred to as "context-dependent". Objects, seen as constellations of local features (interest points, regions, etc.), are matched by minimizing an energy function mixing (1) a fidelity term which measures the quality of feature matching, (2) a neighborhood criteria which captures the object geometry and (3) a regularization term. We will show that the fixed-point of this energy is a "context- dependent" kernel ("CDK") which also satisfies the Mercer condition. Experiments conducted on object recognition show that when plugging our kernel in SVMs, we clearly outperform SVMs with "context-free" kernels.

Domains: Computer Vision and Pattern Recognition [cs.CV]

Hichem Sahbi, Jean-Yves Audibert, Jaonary Rabarisoa, Renaud Keriven. Context-dependent kernel design for object matching and recognition. CVPR, Jun 2008, Anchorage, United States. pp.1-8. ⟨hal-00835081⟩