Active Metric Learning for Object Recognition

  • Sandra Ebert
  • Mario Fritz
  • Bernt Schiele
Part of the Lecture Notes in Computer Science book series (LNCS, volume 7476)


Popular visual representations like SIFT have shown broad applicability across many task. This great generality comes naturally with a lack of specificity when focusing on a particular task or a set of classes. Metric learning approaches have been proposed to tailor general purpose representations to the needs of more specific tasks and have shown strong improvements on visual matching and recognition benchmarks. However, the performance of metric learning depends strongly on the labels that are used for learning. Therefore, we propose to combine metric learning with an active sample selection strategy in order to find labels that are representative for each class as well as improve the class separation of the learnt metric. We analyze several active sample selection strategies in terms of exploration and exploitation trade-offs. Our novel scheme achieves on three different datasets up to 10% improvement of the learned metric. We compare a batch version of our scheme to an interleaved execution of sample selection and metric learning which leads to an overall improvement of up to 23% on challenging datasets for object class recognition.


Active Sampling Label Propagation Sampling Criterion Pairwise Constraint Original Feature Space 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Baram, Y., El-Yaniv, R., Luz, K.: Online Choice of Active Learning Algorithms. JMLR 5, 255–291 (2004)MathSciNetGoogle Scholar
  2. 2.
    Basu, S., Banerjee, A., Mooney, R.: Active Semi-Supervision for Pairwise Constrained Clustering. SIAM (2004)Google Scholar
  3. 3.
    Cebron, N., Berthold, M.R.: Active learning for object classification: from exploration to exploitation. DMKD 18(2), 283–299 (2009)MathSciNetCrossRefGoogle Scholar
  4. 4.
    Chang, C.C., Lin, C.J.: LIBSVM: A library for support vector machines. ACM Transactions on Intelligent Systems and Technology 2(3), 1–27 (2011)Google Scholar
  5. 5.
    Dasgupta, S., Hsu, D.: Hierarchical sampling for active learning. In: ICML (2008)Google Scholar
  6. 6.
    Davis, J., Kulis, B., Jain, P., Sra, S., Dhillon, I.: Information-theoretic metric learning. In: ICML (2007)Google Scholar
  7. 7.
    Ebert, S., Fritz, M., Schiele, B.: Pick Your Neighborhood – Improving Labels and Neighborhood Structure for Label Propagation. In: Mester, R., Felsberg, M. (eds.) DAGM 2011. LNCS, vol. 6835, pp. 152–162. Springer, Heidelberg (2011)CrossRefGoogle Scholar
  8. 8.
    Ebert, S., Fritz, M., Schiele, B.: RALF: A Reinforced Active Learning Formulation for Object Class Recognition. In: CVPR (2012)Google Scholar
  9. 9.
    Goldberger, J., Roweis, S., Hinton, G., Salakhutdinov, R.: Neighbourhood Components Analysis. In: NIPS (2005)Google Scholar
  10. 10.
    Guillaumin, M., Verbeek, J., Schmid, C.: Multiple Instance Metric Learning from Automatically Labeled Bags of Faces. In: Daniilidis, K., Maragos, P., Paragios, N. (eds.) ECCV 2010, Part I. LNCS, vol. 6311, pp. 634–647. Springer, Heidelberg (2010)CrossRefGoogle Scholar
  11. 11.
    Kulis, B., Jain, P., Grauman, K.: Fast Similarity Search for Learned Metrics. PAMI 31(12), 2143–2157 (2009)CrossRefGoogle Scholar
  12. 12.
    Malisiewicz, T., Gupta, A., Efros, A.: Ensemble of Exemplar-SVMs for Object Detection and Beyond. In: ICCV (2011)Google Scholar
  13. 13.
    Nguyen, H.T., Smeulders, A.: Active learning using pre-clustering. In: ICML (2004)Google Scholar
  14. 14.
    Settles, B., Craven, M.: An analysis of active learning strategies for sequence labeling tasks. In: Emp. Meth. in NLP (2008)Google Scholar
  15. 15.
    Straka, M., Hauswiesner, S., Ruether, M., Bischof, H.: Skeletal Graph Based Human Pose Estimation in Real-Time. In: BMVC (2011)Google Scholar
  16. 16.
    Tong, S., Koller, D.: Support Vector Machine Active Learning with Applications to Text Classification. JMLR 2, 45–66 (2001)Google Scholar
  17. 17.
    Vedaldi, A., Fulkerson, B.: VLFEAT: An Open and Portable Library of Computer Vision Algorithms (2008),
  18. 18.
    Yang, L., Jin, R., et al.: Bayesian active distance metric learning. In: UAI (2007)Google Scholar
  19. 19.
    Zhou, D., Bousquet, O., Lal, T.N., Weston, J., Schölkopf, B.: Learning with Local and Global Consistency. In: NIPS (2004)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Sandra Ebert
    • 1
  • Mario Fritz
    • 1
  • Bernt Schiele
    • 1
  1. 1.Max Planck Institute for InformaticsSaarbruckenGermany

Personalised recommendations