Comparative object similarity for improved recognition with few or no examples
Authors:
- 1 June 2010
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 10636919,p. 3525-3532
Abstract
Learning models for recognizing objects with few or no training examples is important, due to the intrinsic long-tailed distribution of objects in the real world. In this paper, we propose an approach to use comparative object similarity. The key insight is that: given a set of object categories which are similar and a set of categories which are dissimilar, a good object model should respond more strongly to examples from similar categories than to examples from dissimilar categories. We develop a regularized kernel machine algorithm to use this category dependent similarity regularization. Our experiments on hundreds of categories show that our method can make significant improvement, especially for categories with no examples.Keywords
This publication has 17 references indexed in Scilit:
- Attribute and simile classifiers for face verification2009
- Max-margin additive classifiers for detection2009
- Describing objects by their attributes2009
- Learning to detect unseen object classes by between-class attribute transfer2009
- A discriminatively trained, multiscale, deformable part model2008
- Representing shape with a spatial pyramid kernel2007
- Beyond Bags of Features: Spatial Pyramid Matching for Recognizing Natural Scene Categories2006
- One-shot learning of object categoriesIEEE Transactions on Pattern Analysis and Machine Intelligence, 2006
- Distinctive Image Features from Scale-Invariant KeypointsInternational Journal of Computer Vision, 2004
- Online Learning with KernelsIEEE Transactions on Signal Processing, 2004