An invariant large margin nearest neighbour classifier

Conference item

The k-nearest neighbour (kNN) rule is a simple and effective method for multi-way classification that is much used in Computer Vision. However, its performance depends heavily on the distance metric being employed. The recently proposed large margin nearest neighbour (LMNN) classifier [21] learns a distance metric for kNN classification and thereby improves its accuracy. Learning involves optimizing a convex problem using semidefinite programming (SDP). We extend the LMNN framework to incorporate knowledge about invariance of the data. The main contributions of our work are three fold: (i) Invariances to multivariate polynomial transformations are incorporated without explicitly adding more training data during learning - these can approximate common transformations such as rotations and affinities; (ii) the incorporation of different regularizes on the parameters being learnt; and (Hi) for all these variations, we show that the distance metric can still be obtained by solving a convex SDP problem. We call the resulting formulation invariant LMNN (lLMNN) classifier. We test our approach to learn a metric for matching (i) feature vectors from the standard Iris dataset; and (ii) faces obtained from TV video (an episode of 'Buffy the Vampire Slayer'). We compare our method with the state of the art classifiers and demonstrate improvements.

Authors: Kumar, MP; Torr, PHS; Zisserman, A

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: IEEE
• Host title: 2007 IEEE 11th International Conference on Computer Vision
• Pages: 1-8
• Publication date: 2007-12-26
• Event title: 2007 IEEE 11th International Conference on Computer Vision (ICCV 2007)
• Event location: Rio de Janeiro, Brazil
• Event start date: 2007-10-14
• Event end date: 2007-10-21
• DOI: 10.1109/iccv.2007.4409041
• EISSN: 2380-7504
• ISSN: 1550-5499
• EISBN: 9781424416318
• ISBN: 9781424416301
• Language: English
• Keywords: testing; face detection; information retrieval; iris; polynomials; TV; computer vision; image recognition; training data; object recognition