An optical character recognition of machine printed Oriya script

Character recognition research is one of the most challenging research areas which put many of the research areas of Digital Image processing like Noise Removal, Machine Learning, and Object Recognition under one roof. If we focus on India particularly, the field has huge amount of work to be done. In current technological era OCR systems for Indian scripts has great application opportunities with their significant use like digital libraries of literatures written in different Indian scripts. Oriya is one of the officially recognized languages of India, and is the official language of the state Orissa. Through this paper I am introducing a Character Recognition System for machine printed Oriya symbols. Paper describes in detail the Feature extraction process, which is carried out by some conventional with two novel enhancements to Structural Features like detection of vertical line in a character and detection of open space in lower zone of a character. Classification process is completed by implementing a novel combination of Binary tree and Naïve Bayesian classifier, which has never been used for Oriya script earlier.