Reconnaissance de l'écriture arabe cursive : combinaison de classifieurs MMCs à fenêtres orientées

Ramy Al-Hajj (1, 2) , Chafic Mokbel (2) , Laurence Likforman-Sulem (1)

1 LTCI - Laboratoire Traitement et Communication de l'Information
2 UOB - University of Balamand [Liban]

The problem addressed in this paper is the recognition of handwritten Arabic city names. Our approach relies on the combination of three HMM-based classifiers. All classifiers have the same right-left topology and use the sliding window approach. The feature set extracted by these classifiers include both baseline independent and baseline dependent features. Each classifier differs from the orientation of the sliding window in order to cope with writing orientation. We combine these classifiers at the decision level and report results on the benchmark IFN/ENIT database of Arabic Tunisian city names.

Domaines: Automatique / Robotique; Traitement du signal et de l'image [eess.SP]; Traitement du signal et de l'image [eess.SP]

Ramy Al-Hajj, Chafic Mokbel, Laurence Likforman-Sulem. Reconnaissance de l'écriture arabe cursive : combinaison de classifieurs MMCs à fenêtres orientées. Sep 2006, pp.271-276. ⟨hal-00111156⟩