Recognition of Handwritten Arabic (Indian) Numerals Using Freeman's Chain Codes and Abductive Network Classifiers

摘要: Accurate automatic recognition of handwritten Arabic numerals has several important applications, e.g. in banking transactions, automation postal services, and other data entry related applications. A number modelling machine learning techniques have been used for recognition, including Neural Network, Support Vector Machine, Hidden Markov Models. This paper proposes the use abductive networks to problem. We studied performance network architecture on a dataset 21120 samples 0-9 digits produced by 44 writers. developed new feature set using histograms contour points chain codes. Recognition rates as high 99.03% were achieved, which surpass reported literature same set. Moreover, technique achieves significant reduction features required.