Multi-scale Gated Fully Convolutional DenseNets for semantic labeling of historical newspaper images

Yann Soullard (1, 2) , Pierrick Tranouez (3) , Clement Chatelain (1, 3) , Stéphane Nicolas (3) , Thierry Paquet (3)

1 LITIS - Laboratoire d'Informatique, de Traitement de l'Information et des Systèmes
2 LETG - Rennes - Littoral, Environnement, Télédétection, Géomatique
3 DocApp - LITIS - Equipe Apprentissage

Domains: Computer Science [cs]; Computer Vision and Pattern Recognition [cs.CV]; Computer Science [cs]; Machine Learning [cs.LG]; Computer Science [cs]; Neural and Evolutionary Computing [cs.NE]; Computer Science [cs]; Document and Text Processing

Yann Soullard, Pierrick Tranouez, Clement Chatelain, Stéphane Nicolas, Thierry Paquet. Multi-scale Gated Fully Convolutional DenseNets for semantic labeling of historical newspaper images. Pattern Recognition Letters, In press, ⟨10.1016/j.patrec.2020.01.026⟩. ⟨hal-02476841⟩