Recognizing cursive Arabic text: Using statistical features and interconnected mono-HMMs

This paper presents a cursive Arabic text recognition system. The system decomposes the document image into text line images and extracts a set of simple statistical features from a one-pixel width window which is sliding a cross that text line. It then injects the resulting feature vectors to Hidden Markov Models. The proposed system is applied to a data corpus which includes Arabic text of more than 600 A4-size sheets typewritten in multiple computer-generated fonts.