Detection of manipulated cheque images in Cheque Truncation System using mismatch in pixels

Abstract

Cheque Truncation System (CTS) is an image based cheque clearing system. It enables faster clearing of cheques. Indian banks have implemented this technology in selected areas of the country. Meanwhile, several fraud cases have already occurred by taking advantage of inadequate security mechanisms of this system. One of several mechanisms of committing frauds is to alter the digital image of the cheque. This paper proposes a method to tackle this type of fraud. It is based on inserting a watermark signal in the image of the cheque. But unlike traditional watermarking, this method is not interested in extracting the watermark at the receiver side. Rather it counts number of mismatching pixels in between the cover image and the watermarked image to detect occurrence of fraud. It also can identify which portions of the cheque image have been altered.