Studies On Off-line Handwritting Arabic Characters Recognition Key Technology

Whether in scientific research or daily life, off-line handwritten character recognition technology has important theoretical significance and broad application prospect,it involves many classic problems in the field of pattern recognition, such as feature extraction, classifier design and samples choice and so on.the research of character recognition problem usually divides into print character recognition and the handwritting character recognition,and the handwritting character recognition can be divided into off-line handwritting character recognition and online handwritting character recognition. Seen from the difficulty of technology,Weaning the handwritting character recognition is more complex and more difficult than online handwritting character recognition,and print text recognition is easier than the handwritting character recognition. Print character recognition technology research and application is already very mature.Whether in Arabic character recognition or other language recognition, it has relatively mature application products.While weaning the handwritting character recognition just has made good grades in the Arabic numeral recognition, recognition effect in Chinese and Arab word is not so good,so needs to do more research to improve the existing circumstances.Research on weaning the script character recognition makes the text information processing more automation,and provides important theoretic meaning and broad application prospect to further enhance the computer intelligent input.Arabic handwriting has its unique charaeteristic and challenges which requires innovative solutions, since existing methods only achieved limited success in constrained cases in this field. In recent years more and more research in statutes started to investigate the recognition of offline handwriting Arabic character. Domestic and international researchers have performed a wide investigation from different perspective in this field,and have achieved some success. However, up until now, there has no commercial produet of offline handwriting Arabic recognition system.The accurate and the choice of testing set of existing system should be improved. It has distinguished distance between the effect and reality. So far, off-line handwritten Arabic character recognition has not yet commercial products available.First,this paper research the key technology of offline handwriting characte recognition,evaluaton of existing related algorithms. And from preprocessing,feature extraction and classifier design etc main aspects analysis this technology. And HMMs Recognition in the character field of the application of a detailed analysis and research.After then,the recognition of cursive handwritting character is still a research focus in the character recognition domain. In this paper we describe an off-line Arabic handwriting recognition system based on semi-continuous hidden markov models of gaussion mixture. The system extracts pixels density and concavity features directly on binary images without complex preprocessing. There are same foreground pixels in the sub-block of each window. The overlap between successive windows is half width of window and the dimension of feature is only 36, these would save the computation of the system. The proposed system builds character models and learns word models using embedded training without character pre-segmentation. Experiments that have been implemented on the benchmark IFN/ENIT database show the average recognition rate of this system with one best output is 86.6%.Finally, the paper concluded were related, and proposes future research priorities and directions.