|
Depot Institutionnel de l'UMBB >
Publications Scientifiques >
Communications Internationales >
Veuillez utiliser cette adresse pour citer ce document :
http://dlibrary.univ-boumerdes.dz:8080/handle/123456789/11260
|
| Titre: | Super-resolution of document images using transfer deep learning of an ESRGAN model |
| Auteur(s): | Kezzoula, Zakia
Gaceb, Djamel
Gritli, Nadjat |
| Mots-clés: | Deep learning
Document images
Image processing
Intelligent vision and AI
Super-resolution |
| Date de publication: | 2022 |
| Editeur: | IEEE |
| Collection/Numéro: | ISIA 2022 - International Symposium on Informatics and its Applications, Proceedings;pp. 1-6 |
| Résumé: | This paper presents a novel super-resolution approach of document images. It is based on transfer deep learning of an ESRGAN model. This model, which showed good robustness on natural images, has been adapted to document images by using better levels of fine-tuning and a post-processing to enhance contrast. The experiments were carried out on our document image dataset that we built from document images presenting different challenges. Documents of different categories with different complexity levels and degradation kinds. The results obtained are better compared to ten existing approaches, which we have developed and tested on the same dataset with the same evaluation protocol |
| URI/URL: | DOI 10.1109/ISIA55826.2022.9993497
https://ieeexplore.ieee.org/document/9993497
http://dlibrary.univ-boumerdes.dz:8080/handle/123456789/11260 |
| ISBN: | 978-166547473-3 |
| Collection(s) : | Communications Internationales
|
Fichier(s) constituant ce document :
Il n'y a pas de fichiers associés à ce document.
|
Tous les documents dans DSpace sont protégés par copyright, avec tous droits réservés.
|
