A Bank Card Number Recognition Method Based on Convolutional Neural Network

With the popularity of online transactions and big data platforms, bank card digital recognition can save people from manually entering a long list of card numbers. Traditional optical way and machine learning methods have low identification accuracy and are susceptible to background environment interference. To solve this problem, this paper proposes a bank card number recognition method, which is a cooperation of convolutional neural network and character positioning. Firstly, various bank card pictures shot from different angles are collected as our input data; then EAST network algorithm is used to realize character positioning; finally, the image after character positioning segmentation is input into the convolutional neural network and vgg16 is used for transfer learning to realize card number recognition. Experimental results show that this method has better recognition accuracy and robustness.