Design and Implementation of Stroke Based Hand Written Character Recognition

Abstract

Computerization in almost every industry and field helping to speed up the process and maintaining the record in very organized way. Still the computerization is not feasible to replace every process in different job profile. Secondly, computerization is feasible for afterward process but still it need to convert existing information or the documents in electronics format for further processing. From the doctors point of view it is bit difficult to handle patients with their complaints or diagnostics and operate the computer or electronic gadgets to digitally record or generate report is complicated or the cumbersome. Biggest challenge in medical field is doctor’s prescription writing style which is difficult to understand and hence again difficult for OCR (Optical Character Recognition) like software to recognize the written characters. OCR usually utilized pattern-matching algorithm for detection process. After study, it found pattern matching is not feasible for badly hand written character. Human brain understands character based on writing strokes of character and this allow humans to read character in different font also. This paper research mainly target at development of stroke-based hand written character recognition.