Title :
A neuro-fuzzy approach to recognize Arabic handwritten characters
Author_Institution :
Dept. of Electr. Eng., Ecole Nationale d´´Ingenieurs de Sfax, Tunisia
Abstract :
In this paper we describe a system that recognizes online Arabic handwritten characters. In this system, a fuzzy neural network is used to classify characters. The characters used in this system were segmented from cursive handwriting that are modelled by a theory of movement generation. Based on this theory, the features extracted from each character are the neuro-physiological parameters of the equation describing the curvilinear velocity of the script. For each character presented to the system, a fuzzy membership is assigned to each output of the neural network
Keywords :
character recognition; feature extraction; fuzzy neural nets; real-time systems; Arabic handwritten character recognition; cursive handwriting; curvilinear velocity; feature extraction; fuzzy membership; fuzzy neural network; online character recognition; segmentation; Character generation; Character recognition; Equations; Feature extraction; Fuzzy neural networks; Handwriting recognition; Neural networks; Pattern recognition; Shape; Tail;
Conference_Titel :
Neural Networks,1997., International Conference on
Conference_Location :
Houston, TX
Print_ISBN :
0-7803-4122-8
DOI :
10.1109/ICNN.1997.613998
