User menu

Accès à distance ? S'identifier sur le proxy UCLouvain

Training with corrupted labels to reinforce a probably correct teamsport player detector

Primary tabs

Parisot, Pascaline [UCL] Sevilmis, Berk [UCL] De Vleeschouwer, Christophe [UCL]

While the analysis of foreground silhouettes has become a key component of modern approach to multi-view people detection, it remains subject to errors when dealing with a single viewpoint. Besides, several works have demonstrated the benefit of exploiting classifiers to detect objects or people in images, based on local texture statistics. In this paper, we train a classifier to differentiate false and true positives among the detections computed based on a foreground mask analysis. This is done in a sport analysis context where people deformations are important, which makes it important to adapt the classifier to the case at hand, so as to take the teamsport color and the background appearance into account. To circumvent the manual annotation burden incurred by the repetition of the training for each event, we propose to train the classifier based on the foreground detector decisions. Hence, since the detector is not perfect, we face a training set whose labels might be corrupted. We investigate a set of classifier design strategies, and demonstrate the effectiveness of the approach to reliably detect sport players with a single view.

Document type Communication à un colloque (Conference Paper) – Présentation orale avec comité de sélection
Access type Accès libre
Publication date 2013
Language Anglais
Conference "Advanced Concepts for Intelligent Vision Systems", Poznan (Poland) (du 28/10/2013 au 31/10/2013)
Peer reviewed yes
Host document "Advanced Concepts for Intelligent Vision Systems, Poznan, Poland, Lecture Notes in Computer Science Volume 8192" (ISBN : 978-3-319-02895-8)
Publication status Publié
Affiliation UCL - SST/ICTM/ELEN - Pôle en ingénierie électrique
Keywords detection ; random ferns ; corrupted label
Links
  1. Alahi, A., Jacques, L., Boursier, Y., Vandergheynst, P.: Sparsity driven people localization with a heterogeneous network of cameras. Jour. of MIV 41(1-2), 39–58 (2011)
  2. Amit Yali, Geman Donald, Shape Quantization and Recognition with Randomized Trees, 10.1162/neco.1997.9.7.1545
  3. Blum Avrim, Mitchell Tom, Combining labeled and unlabeled data with co-training, 10.1145/279943.279962
  4. Bosch Anna, Zisserman Andrew, Munoz Xavier, Image Classification using Random Forests and Ferns, 10.1109/iccv.2007.4409066
  5. Chen Fan, Delannay Damien, De Vleeschouwer Christophe, An Autonomous Framework to Produce and Distribute Personalized Team-Sport Video Summaries: A Basketball Case Study, 10.1109/tmm.2011.2166379
  6. Delannay, D., Danhier, N., De Vleeschouwer, C.: Detection and recognition of sports (wo)men from multiple views. In: Proc. of ACM/IEEE ICDSC (2009)
  7. Dollar Piotr, Tu Zhuowen, Perona Pietro, Belongie Serge, Integral Channel Features, 10.5244/c.23.91
  8. Dollar P., Wojek C., Schiele B., Perona P., Pedestrian detection: A benchmark, 10.1109/cvpr.2009.5206631
  9. Felzenszwalb P F, Girshick R B, McAllester D, Ramanan D, Object Detection with Discriminatively Trained Part-Based Models, 10.1109/tpami.2009.167
  10. Fleuret F., Berclaz J., Lengagne R., Fua P., Multicamera People Tracking with a Probabilistic Occupancy Map, 10.1109/tpami.2007.1174
  11. Freund, Y., Schapire, R.E.: A decision-theoretic generalization of on-line learning and an application to boosting. Jour. of CSS 55(1), 119–139 (1997)
  12. Geurts Pierre, Ernst Damien, Wehenkel Louis, Extremely randomized trees, 10.1007/s10994-006-6226-1
  13. Khan Saad M., Shah Mubarak, A Multiview Approach to Tracking People in Crowded Scenes Using a Planar Homography Constraint, Computer Vision – ECCV 2006 (2006) ISBN:9783540338383 p.133-146, 10.1007/11744085_11
  14. Levin, A., Viola, P., Freund, Y.: Unsupervised improvement of visual detectors using co-training. In: ICCV, pp. 626–633 (2003)
  15. Maree R., Geurts P., Piater J., Wehenkel L., Random Subwindows for Robust Image Classification, 10.1109/cvpr.2005.287
  16. Nair V., Clark J.J., An unsupervised, online learning framework for moving object detection, 10.1109/cvpr.2004.1315181
  17. Ozuysal M., Calonder M., Lepetit V., Fua P., Fast Keypoint Recognition Using Random Ferns, 10.1109/tpami.2009.23
  18. Roth Peter, Grabner Helmut, Skočaj Danijel, Bischof Horst, Leonardis Aleš, Conservative Visual Learning for Object Detection with Minimal Hand Labeling Effort, Lecture Notes in Computer Science (2005) ISBN:9783540287032 p.293-300, 10.1007/11550518_37
  19. Viola, P., Jones, M.: Robust real-time object detection. In: Proc. of the Int. Workshop on SCTV (2001)
  20. Junliang Xing, Haizhou Ai, Liwei Liu, Shihong Lao, Multiple Player Tracking in Sports Video: A Dual-Mode Two-Way Bayesian Inference Approach With Progressive Observation Modeling, 10.1109/tip.2010.2102045
Bibliographic reference Parisot, Pascaline ; Sevilmis, Berk ; De Vleeschouwer, Christophe. Training with corrupted labels to reinforce a probably correct teamsport player detector.Advanced Concepts for Intelligent Vision Systems (Poznan (Poland), du 28/10/2013 au 31/10/2013). In: Advanced Concepts for Intelligent Vision Systems, Poznan, Poland, Lecture Notes in Computer Science Volume 8192, 2013
Permanent URL http://hdl.handle.net/2078.1/133326