Odia character recognition system: A study on feature extraction and classification techniques

Abstract Optical Character Recognition (OCR) is the most successful application for automatic pattern recognition. OCR has been a widely studied field of research and development for more than last 50 years. It is one of the many important gifts given by computer science to the mankind. OCR means “a technique of recognition of machine printed or hand written text by computer and its conversion to an editable form as per the requirement”. During the recent decades, handwritten character recognition has been a subject of study because of both its theoretical importance in recognition system with numerous possible applications. This paper presents an extensive study of numerous publications by a number of researchers working for years in the field of Odia character recognition (one of the many regional languages used in India and spoken by more than 30 million people worldwide). The main focus has been on feature extraction, data set size, classification techniques and recognition accuracy. A number of selective articles published in journals and conference proceedings on character recognition have been included in this work (2005 to 2020). The review process has been divided into two parts. In the first part a study of various works by researchers is presented with detailed description. In the second part, various feature extraction and classification techniques have been presented and summarized. The findings are presented in the analysis section.