OCR to read embossed text from Credit/Debit card
Pranav Raka, Sakshi Agrawal · International journal of advance research, ideas and innovations in technology · 2019
This paper is based on a real application in which the task was to recognize the digits and characters embossed on the credit/debit card i.e the card number and the expiry date. The problem was that the digit embossed are uneven. Using tesseract for OCR was not enough to recognize the digits accurately. The first problem was that to separate the digits and the second fact was that the separated characters or digits were distorted heavily. Challenge is to recognize the digits correctly. In this paper, we present an algorithm for credit/debit card number identification based on Optical Character Recognition using KNN. The algorithm is tested on five hundred credit and debit card images of different illumination of lights. When the images have captured the angle of view and distance were taken into consideration. The images were then preprocessed using an image processing algorithm. The resultant images were then fed to the OCR system. The achievement of this work was the correct identification of digits with 90% and zero false identification.