Designing Offline Arabic Handwritten Isolated Character Recognition System using Artificial Neural Network Approach

The Arabic language is one of the major languages that has little attention in character recognition field by Arab researchers in particular and foreign researchers in general. Due to the highly cursive nature of handwritten Arabic language, Arabic character recognition is considered one of the most challenging problems in contrast to working with Latin, Japanese or Chinese character recognition. In this paper, we proposed Arabic off-line handwritten isolated recognition system based on novel feature extraction techniques, a back propagation artificial neural network as classification phase. The presented work is implemented and tested via the CENPARMI database. Competitive recognition accuracy has been achieved 96.14%. This result motivates us and other researchers in this field to employ the features extraction techniques that we have used in this research with other Arabic character shapes.

The Arabic language is one of the major languages that has little attention in character recognition field by Arab researchers in particular and foreign researchers in general. Due to the highly cursive nature of handwritten Arabic language, Arabic character recognition is considered one of the most challenging problems in contrast to working with Latin, Japanese or Chinese character recognition. In this paper, we proposed Arabic off-line handwritten isolated recognition system based on novel feature extraction techniques, a back propagation artificial neural network as classification phase. The presented work is implemented and tested via the CENPARMI database. Competitive recognition accuracy has been achieved 96.14%. This result motivates us and other researchers in this field to employ the features extraction techniques that we have used in this research with other Arabic character shapes.