MICR Automated Recognition based on Paraconsistent Artificial Neural Networks
Sheila Souza, Jair Minoro Abe, Kazumi Nakamatsu · Procedia Computer Science · 2013
The purpose of this paper is to discuss an automated computational system able to recognize MICR characters commonly used on bank checks based on Paraconsistent Artificial Neural Networks due to their intrinsic ability to deal with imprecise, inconsistent and paracomplete data. The recognition process is carried out from character features chosen in advance based on Graphology and Graphoscopy techniques. The analysis of such features and the character recognition are performed employing Paraconsistent Artificial Neural Networks. Actual checks batches were presented to validate the proposed study and 97.8 percent of the characters were recognized correctly by the system.