|Semester Project, EPFL||June 2009|
The purpose of this work is to develop a system for improving the percentage of successful of the optical character recognition softwares (OCR). OCR softwares generate text files from an image of printed or typed text which can be recognized/exploited by text editing programs.
Used on a raw image, these programs may lack strength and have an unsatisfactory percentage of successful.
The idea is to process the images before using the OCR in order to improve efficiency. Here is the result of such image processing.