Design & implementation of a novel cognitive character recognition technique

In today´s competitive world electronic document management systems are of great benefit to society, but the paper document has not disappeared from the actual business sites because of its excellent characteristics. The process of analyzing and recognizing documents involves various steps. An important step in this process is automatic recognition of characters from the document. Character recognition being one of the main areas in the field of computer sciences is still an open area for researchers. Character recognition is a trivial task for humans, but the development of its computer based solution is extremely difficult because of many sources of variability. Number of algorithms have been devised and implemented which recognize the characters by exploiting different characteristics of the characters input to them. Humans achieve instantaneous recognition of objects without contextual analysis, which is primarily accomplished by recognition in a glance the appearance of object and observing strokes & relationships between them. This paper explores one such algorithm which is very close to human recognition system of objects and attempts to recognize characters by analyzing minimal amount of data. The method proposed is independent of the size of the character being recognized and is able to accommodate character styles which have slight variation.