论文信息 - UPMC-LIP6 at TrecVid'08: Balanced and Unbalanced Forests of Fuzzy Decision Trees for High-Level Feature Detection

UPMC-LIP6 at TrecVid'08: Balanced and Unbalanced Forests of Fuzzy Decision Trees for High-Level Feature Detection

In this paper, we present the methodology we applied in our submission to the NIST TRECVID’2008 evaluation. We participated in the High-level Feature Extraction task. Our approach is based on the use of a Forest of Fuzzy Decision Trees constructed by means of either a balanced sampling of the development data, or an unbalanced sampling of the development data.

Christophe Marsala | Marcin Detyniecki

[1] Leo Breiman,et al. Random Forests , 2001, Machine Learning.

[2] Leo Breiman,et al. Bagging Predictors , 1996, Machine Learning.

[3] Christophe Marsala,et al. University of Paris 6 at TRECVID 2005: High-Level Feature Extraction , 2005, TRECVID.

[4] Massih-Reza Amini,et al. UHigh-Level Feature Detection with Forests of Fuzzy Decision Trees combined with the RankBoost Algorithm , 2007, TRECVID.

[5] C. Marsala,et al. Apprentissage inductif en presence de donnees imprecises : construction et utilisation d'arbres de decision flous , 1998 .

[6] Christophe Marsala,et al. University of Paris 6 at TRECVID 2006: Forest of Fuzzy Decision Trees for High-Level Feature Extraction , 2006, TRECVID.

[7] Massih-Reza Amini,et al. High-Level Feature Detection with Forests of Fuzzy Decision Trees combined with the RankBoost , 2007 .

[8] Paul Over,et al. Evaluation campaigns and TRECVid , 2006, MIR '06.

[9] Christian Petersohn. Fraunhofer HHI at TRECVID 2004: Shot Boundary Detection System , 2004, TRECVID.