OCR Post Processing Based on Character Pattern Matching (Optical Character Recognition) (Report‪)‬

1. INTRODUCTION The domain of digital image processing is in a continuous expansion and for this reason more and more solutions appear to different problems. However, each problem solved creates a new possible improvement and so new methods are developed all the time. Digital image processing can be divided into a large number of sub categories that deal with different aspects of a digital image. There are algorithms dealing with different conversions from one colour scheme to another. Other algorithms deal with the recognition of different elements found in an image. The goal of this last type of algorithms is to simulate the human eye by matching different patterns with already known models. Matching patterns can vary from simple lines to more complicated characters that contain a variety of geometrical shapes.