Research On Credit Card Identification Based On Mobile Phone

With the rapid development of ecommerce and smart phones in China, more and more people are going to do online shopping by mobile device. During the payment period, usually, the storefront will ask users to enter the credit card number, but it will be hard for the user to remember all the credit card digits and most likely users will make mistake when entering the card numbers. So, it will be a headache for users to do online shopping. Some storefronts are going to improve the user shopping experience by asking user to take picture for the credit card and let the mobile device recognize the card numbers. But, usually, the credit card had a complex background, and the credit card numbers will have a similar contrast as the background. So, it will be hard to recognize this kind of raised character.The credit card number detector will consist of the hardware and the software. In this thesis, in the first step, it will detect the credit card range in the whole picture using2D wavelet transfer. In the second step, it will detect the credit card number characters using the location information, since card number has a fixed location on the credit card. In the third step, it will split the credit card characters and then use the gabor filter to get the single character feature. In the fourth step, we will do the PCA to reduce the data and then use SVM to recognize the single character. Finally, we will verify the credit card numbers by using Luhn algorithm.