Orientation covariant aggregation of local descriptors with embeddings

Giorgos Tolias 1, * Teddy Furon 1 Hervé Jégou 1
* Auteur correspondant
1 TEXMEX - Multimedia content-based indexing
IRISA - Institut de Recherche en Informatique et Systèmes Aléatoires, Inria Rennes – Bretagne Atlantique
Abstract : Image search systems based on local descriptors typically achieve orientation invariance by aligning the patches on their dominant orientations. Albeit successful, this choice introduces too much invariance because it does not guarantee that the patches are rotated consistently. This paper introduces an aggregation strategy of local descriptors that achieves this covariance property by jointly encoding the angle in the aggregation stage in a continuous manner. It is combined with an efficient monomial embedding to provide a codebook-free method to aggregate local descriptors into a single vector representation. Our strategy is also compatible and employed with several popular encoding methods, in particular bag-of-words, VLAD and the Fisher vector. Our geometric-aware aggregation strategy is effective for image search, as shown by experiments performed on standard benchmarks for image and particular object retrieval, namely Holidays and Oxford buildings.
Type de document :
Communication dans un congrès
European Conference on Computer Vision, Sep 2014, Zurich, Switzerland. 2014
Liste complète des métadonnées

Littérature citée [43 références]  Voir  Masquer  Télécharger


https://hal.inria.fr/hal-01020823
Contributeur : Giorgos Tolias <>
Soumis le : mardi 25 novembre 2014 - 10:40:30
Dernière modification le : mercredi 11 avril 2018 - 02:00:46
Document(s) archivé(s) le : jeudi 26 février 2015 - 11:01:06

Fichiers

eccv14_tolias_orientation_cova...
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01020823, version 3
  • ARXIV : 1407.2170

Citation

Giorgos Tolias, Teddy Furon, Hervé Jégou. Orientation covariant aggregation of local descriptors with embeddings. European Conference on Computer Vision, Sep 2014, Zurich, Switzerland. 2014. 〈hal-01020823v3〉

Partager

Métriques

Consultations de la notice

281

Téléchargements de fichiers

851