Video-Based Content Extraction Algorithm from Bank Cards for iOS Mobile Devices

Abstract

This paper proposes an algorithm for information fields detection and recognition of bank cards in video sequences, obtained from the iOS mobile device camera. For this, we use two basic steps. The first step is preprocessing for localization of symbols. The second step is symbol block recognition using OCR system. Preprocessing algorithm includes card edge detection, information fields segmentation, segments enhancement, symbols edge emphasizing and the final step is symbol block recognition using OCR system. Based on our approach and iPhone SDK frameworks, OpenCV and Tesseract library the bank card details recognition software is implemented. For experiments our database of the real static images and video sequences was used.