Extraction and Organization of the Useful Signal of Arabic Off Line Handwritten Tracing

In this paper, we present a system of extraction and organization of the useful signal of arabic off line handwritten tracing. The calculation of two curvature functions permits to extract the dominant points on the contour of the word. An appropriated algorithm permits to extract the useful signal of the stroke that represents a succession of the points belonging to its median axis. Finally, another algorithm permits to detect the starting point and to follow the useful signal extracted in order to reconstitute the order in which it has been written. To validate this system, we tested it on several words of the data base of Tunisian city names IFN/ENIT