IRIM at TRECVID2009: High Level Feature ExtractionReport as inadecuate




IRIM at TRECVID2009: High Level Feature Extraction - Download this document for free, or read online. Document in PDF available to download.

1 LIST - Laboratoire d-Intégration des Systèmes et des Technologies 2 ETIS - Equipe traitement des images et du signal 3 Eurecom Sophia Antipolis 4 MIDI ETIS - Equipe traitement des images et du signal 5 GIPSA-Services - GIPSA-Services GIPSA-lab - Grenoble Images Parole Signal Automatique 6 GIPSA-GPIG - GPIG GIPSA-DIS - Département Images et Signal 7 IRIT - Institut de recherche en informatique de Toulouse 8 LaBRI - Laboratoire Bordelais de Recherche en Informatique 9 LEAR - Learning and recognition in vision Inria Grenoble - Rhône-Alpes, LJK - Laboratoire Jean Kuntzmann, INPG - Institut National Polytechnique de Grenoble 10 LIF - Laboratoire d-informatique Fondamentale de Marseille - UMR 6166 11 LIG - Laboratoire d-Informatique de Grenoble 12 MALIRE - Machine Learning and Information Retrieval LIP6 - Laboratoire d-Informatique de Paris 6 13 LSIS - Laboratoire des Sciences de l-Information et des Systèmes 14 SIC XLIM - XLIM, Université de Poitiers

Abstract : The IRIM group is a consortium of French teams working on Multimedia Indexing and Retrieval. This paper describes our participation to the TRECVID 2009 High Level Features detection task. We evaluated a large number of different descriptors on TRECVID 2008 data and tried different fusion strategies, in particular hierarchical fusion and genetic fusion. The best IRIM run has a Mean Inferred Average Precision of 0.1220, which is significantly above TRECVID 2009 HLF detection task median performance. We found that fusion of the classification scores from different classifier types improves the performance and that even with a quite low individual performance, audio descriptors can help.

Keywords : descriptors high level features extraction fusion classification





Author: Bertrand Delezoide - Hervé Le Borgne - Pierre-Alain Moëllic - David Gorisse - Frédéric Precioso - Feng Wang - Bernard Meriald

Source: https://hal.archives-ouvertes.fr/



DOWNLOAD PDF




Related documents