Exploiting Photographic Style for Category-Level Image Classification by Generalizing the Spatial Pyramid

Gemert Jan C. Van 1
1 WILLOW - Models of visual object recognition and scene understanding
DI-ENS - Département d'informatique de l'École normale supérieure, Inria Paris-Rocquencourt, CNRS - Centre National de la Recherche Scientifique : UMR8548
Abstract : This paper investigates the use of photographic style for category-level image classi cation. Speci cally, we exploit the assumption that images within a category share a similar style de ned by attributes such as colorfulness, lighting, depth of eld, viewpoint and saliency. For these style attributes we create correspondences across images by a generalized spatial pyramid matching scheme. Where the spatial pyramid groups features spatially, we allow more general feature grouping and in this paper we focus on grouping images on photographic style. We evaluate our approach in an object classi cation task and investigate style differences between professional and amateur photographs. We show that a generalized pyramid with style-based attributes improves performance on the professional Corel and amateur Pascal VOC 2009 image datasets.
Type de document :
Communication dans un congrès
ICMR '11 - International Conference on Multimedia Retrieval, Apr 2011, Trento, Italy. ACM, 2011, 〈10.1145/1991996.1992010〉
Liste complète des métadonnées

Littérature citée [24 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01063326
Contributeur : Suha Kwak <>
Soumis le : jeudi 11 septembre 2014 - 17:49:42
Dernière modification le : mardi 24 avril 2018 - 17:20:13
Document(s) archivé(s) le : vendredi 12 décembre 2014 - 10:57:04

Fichier

gemertICMR11.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Collections

Citation

Gemert Jan C. Van. Exploiting Photographic Style for Category-Level Image Classification by Generalizing the Spatial Pyramid. ICMR '11 - International Conference on Multimedia Retrieval, Apr 2011, Trento, Italy. ACM, 2011, 〈10.1145/1991996.1992010〉. 〈hal-01063326〉

Partager

Métriques

Consultations de la notice

428

Téléchargements de fichiers

178