Text-Edge-Box: An Object Proposal Approach for Scene Text Localization - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2017

Text-Edge-Box: An Object Proposal Approach for Scene Text Localization

Dinh Nguyen Van
  • Fonction : Auteur correspondant
  • PersonId : 1014588

Connectez-vous pour contacter l'auteur
Lu Shijian
  • Fonction : Directeur scientifique
Ouarti Nizar
  • Fonction : Auteur

Résumé

Text proposal has been gaining interest in recent years due to the great success of object proposal in categories-independent object localization. In this paper, we present a novel text-specific proposal technique that provides superior bounding boxes for accurate text localization in scenes. The proposed technique, which we call Text Edge Box (TEB), uses a binary edge map, a gradient map and an orientation map of an image as inputs. Connected components are first found within the binary edge map, which are scored by two proposed low-cue text features that are extracted in the gradient map and the orientation map, respectively. These scores present text probability of connected components and are aggregated in a text edge image. Scene texts proposals are finally generated by grouping the connected components and estimating their likelihood of being words. The proposed TEB has been evaluated on the two public scene text datasets: the Robust Reading Competition 2013 dataset (ICDAR 2013) dataset and the Street View Text (SVT) dataset. Experiments show that the proposed TEB outperforms the state-of-the-art techniques greatly.
Fichier non déposé

Dates et versions

hal-01502710 , version 1 (06-04-2017)

Identifiants

Citer

Dinh Nguyen Van, Lu Shijian, Ouarti Nizar, Mounir Mokhtari. Text-Edge-Box: An Object Proposal Approach for Scene Text Localization. Winter Conference on Application and Computer Vision 2017, Mar 2017, Santa Rosa, United States. ⟨10.1109/WACV.2017.149⟩. ⟨hal-01502710⟩
213 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More