A Comparative Study of Diversity Methods for Hybrid Text and Image Retrieval Approaches

Sabrina Tollari (1) , Philippe Mulhem (2) , Marin Ferecatu (3) , Hervé Glotin (4) , Marcin Detyniecki (1) , Patrick Gallinari (1) , Hichem Sahbi (5) , Zhong-Qiu Zhao

1 MALIRE - Machine Learning and Information Retrieval
2 LGI - IMAG - Laboratoire de Génie Informatique
3 TSI - Département Traitement du Signal et des Images
4 LSIS - Laboratoire des Sciences de l'Information et des Systèmes
5 LTCI - Laboratoire Traitement et Communication de l'Information

This article compares eight different diversity methods: 3 based on visual information, 1 based on date information, 3 adapted to each topic based on location and visual information; finally, for completeness, 1 based on random permutation. To compare the effectiveness of these methods, we apply them on 26 runs obtained with varied methods from different research teams and based on different modalities. We then discuss the results of the more than 200 obtained runs. The results show that query-adapted methods are more efficient than non-adapted method, that visual only runs are more difficult to diversify than text only and text-image runs, and finally that only few methods maximize both the precision and the cluster recall at 20 documents.

Domaines: Apprentissage [cs.LG]; Recherche d'information [cs.IR]

Sabrina Tollari, Philippe Mulhem, Marin Ferecatu, Hervé Glotin, Marcin Detyniecki, et al.. A Comparative Study of Diversity Methods for Hybrid Text and Image Retrieval Approaches. 9th Workshop of the Cross-Language Evaluation Forum, CLEF 2008, Sep 2008, Aarhus, Denmark. pp.585-592, ⟨10.1007/978-3-642-04447-2_72⟩. ⟨hal-00581664⟩