Handwritten Word Spotting with Corrected Attributes

Jon Almazan 1, * Albert Gordo 2, * Alicia Fornés 1 Ernest Valveny 1
* Auteur correspondant
2 LEAR - Learning and recognition in vision
Inria Grenoble - Rhône-Alpes, LJK - Laboratoire Jean Kuntzmann, INPG - Institut National Polytechnique de Grenoble
Abstract : We propose an approach to multi-writer word spotting, where the goal is to find a query word in a dataset comprised of document images. We propose an attributes-based approach that leads to a low-dimensional, fixed-length representation of the word images that is fast to compute and, especially, fast to compare. This approach naturally leads to an unified representation of word images and strings, which seamlessly allows one to indistinctly perform query-by-example, where the query is an image, and query-by-string, where the query is a string. We also propose a calibration scheme to correct the attributes scores based on Canonical Correlation Analysis that greatly improves the results on a challenging dataset. We test our approach on two public datasets showing state-of-the-art results.
Type de document :
Communication dans un congrès
ICCV - IEEE International Conference on Computer Vision, Dec 2013, Sydney, Australia. IEEE, pp.1017-1024, 2013, 〈10.1109/ICCV.2013.130〉
Liste complète des métadonnées

Littérature citée [30 références]  Voir  Masquer  Télécharger


https://hal.inria.fr/hal-00906787
Contributeur : Thoth Team <>
Soumis le : mercredi 20 novembre 2013 - 12:48:15
Dernière modification le : jeudi 11 janvier 2018 - 06:21:55
Document(s) archivé(s) le : vendredi 21 février 2014 - 04:28:11

Fichiers

handwritten_iccv13.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Collections

Citation

Jon Almazan, Albert Gordo, Alicia Fornés, Ernest Valveny. Handwritten Word Spotting with Corrected Attributes. ICCV - IEEE International Conference on Computer Vision, Dec 2013, Sydney, Australia. IEEE, pp.1017-1024, 2013, 〈10.1109/ICCV.2013.130〉. 〈hal-00906787〉

Partager

Métriques

Consultations de la notice

420

Téléchargements de fichiers

564