Affordable Access

Access to the full text

Learning to Recognize Objects with Little Supervision

Authors
  • Carbonetto, Peter1
  • Dorkó, Gyuri2
  • Schmid, Cordelia2
  • Kück, Hendrik1
  • de Freitas, Nando1
  • 1 University of British Columbia, Vancouver, Canada , Vancouver (Canada)
  • 2 INRIA Rhône-Alpes, Grenoble, France , Grenoble (France)
Type
Published Article
Journal
International Journal of Computer Vision
Publisher
Springer-Verlag
Publication Date
Jul 21, 2007
Volume
77
Issue
1-3
Pages
219–237
Identifiers
DOI: 10.1007/s11263-007-0067-7
Source
Springer Nature
Keywords
License
Yellow

Abstract

This paper shows (i) improvements over state-of-the-art local feature recognition systems, (ii) how to formulate principled models for automatic local feature selection in object class recognition when there is little supervised data, and (iii) how to formulate sensible spatial image context models using a conditional random field for integrating local features and segmentation cues (superpixels). By adopting sparse kernel methods, Bayesian learning techniques and data association with constraints, the proposed model identifies the most relevant sets of local features for recognizing object classes, achieves performance comparable to the fully supervised setting, and obtains excellent results for image classification.

Report this publication

Statistics

Seen <100 times