Modélisation de HMM en contexte avec des arbres de décision pour la reconnaissance de mots manuscrits

This paper presents an HMM-based recognizer for the off-line recognition of handwritten words. Word models are the concatenation of context-dependent character models: the trigraphs. Due to the large number of possible context-dependent models to compute, a clustering is applied on each state position, based on decision trees. Our system is shown to perform better than a baseline context independent system, and reaches an accuracy higher than 80% on the publicly available Rimes database.