AXES at TRECVid 2012: KIS, INS, and MED - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Document Associé À Des Manifestations Scientifiques Année : 2012

AXES at TRECVid 2012: KIS, INS, and MED

Dan Oneata
  • Fonction : Auteur
  • PersonId : 946916
Jérôme Revaud
  • Fonction : Auteur
  • PersonId : 946914
Zaid Harchaoui
  • Fonction : Auteur
  • PersonId : 895242
Jakob Verbeek
Cordelia Schmid
  • Fonction : Auteur
  • PersonId : 831154
Robin Aly
  • Fonction : Auteur
Relja Arandjelovic

Résumé

The AXES project participated in the interactive instance search task (INS), the known-item search task (KIS), and the multimedia event detection task (MED) for TRECVid 2012. As in our TRECVid 2011 system, we used nearly identical search systems and user interfaces for both INS and KIS. Our interactive INS and KIS systems focused this year on using classifiers trained at query time with positive examples collected from external search engines. Participants in our KIS experiments were media professionals from the BBC; our INS experiments were carried out by students and researchers at Dublin City University. We performed comparatively well in both experiments. Our best KIS run found 13 of the 25 topics, and our best INS runs outperformed all other submitted runs in terms of P@100. For MED, the system presented was based on a minimal number of low-level descriptors, which we chose to be as large as computationally feasible. These descriptors are aggregated to produce high-dimensional video-level signatures, which are used to train a set of linear classifiers. Our MED system achieved the second-best score of all submitted runs in the main track, and best score in the ad-hoc track, suggesting that a simple system based on state-of-the-art low-level descriptors can give relatively high performance. This paper describes in detail our KIS, INS, and MED systems and the results and findings of our experiments.
Fichier principal
Vignette du fichier
oneata12tv.pdf (8.18 Mo) Télécharger le fichier
Vignette du fichier
Selection_001.png (105.64 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Format : Figure, Image
Loading...

Dates et versions

hal-00746874 , version 1 (29-10-2013)
hal-00746874 , version 2 (07-11-2013)

Identifiants

  • HAL Id : hal-00746874 , version 2

Citer

Dan Oneata, Matthijs Douze, Jérôme Revaud, Schwenninger Jochen, Danila Potapov, et al.. AXES at TRECVid 2012: KIS, INS, and MED. TRECVID Workshop, Sep 2012, Gaithersburg, United States. ⟨hal-00746874v2⟩
856 Consultations
1048 Téléchargements

Partager

Gmail Facebook X LinkedIn More