On-line arabic handwriting segmentation via perceptual codes: application to MAYASTROUN database

In this paper, we propose a new approach for Arabic on-line handwriting segmentation via perceptual codes. To identify patterns, our human perceptual system is based on basic features called perceptual codes. Analysing handwriting, we notice the existence of elementary and global ones. Perceptual organization of the elementary perceptual codes in various constraints we obtain global ones, and to obtain different form of handwriting, we proceed by gathering them. So, we present a new approach for on-line handwriting with perceptual codes. This system use the fuzzy set theory to detect the elementary perceptual codes (EPCs) and the genetic algorithms for the global perceptual ones (GPCs). To validate our new approach, MAYASTROUN database was developed containing unconstrained large Arabic lexicon: words alphabet and digit. This database is acquired by different writers on digitizing tablet and indexed by XML. The obtained results show successful representations of Arabic handwritten script via perceptual codes.