Defending FinTech: A Novel Approach to Credit Card Validation using Binary Pattern Features in Machine Learning

Fraud is misappropriating a profit organization’s system that does not always result in immediate legal penalties. Credit card firms must be able to spot fraudulent credit card transactions so that customers aren’t charged for goods they didn’t buy. Nowadays, misusing of credit cards is increasing day by day. In order to stop these illegal activities, a method has been developed in which credit cards are used fraudulently. The method proposed, local binary pattern (LBP), Local Ternary Pattern (LTP), and Local Gradient Pattern (LGP) for characters and numerals features to train the network utilized to detect fraud in credit cards. By combining all these three patterns, size and space between character and numerical are efficiently validated. In the introduced method 70 percent of the image from the standard database has been used for training and 30 percent test image for testing the trained network. The results of the proposed method achieve better than the existing methods interms of accuracy and 92.24% accuracy is achieved through this proposed method.