An Improved Design and Implementation of Chinese Name Card Recognition System on Mobile Devices

Abstract

Since name card recognition based on built-in cameras has become popular people now pay more attention to the robustness and effectiveness of the system. In spite of the maturity of OCR some bottlenecks in lower processing level such as spatial correction and segmentation weak its power. Traditional models treating the card image as an orthogonal projection with only affine distortion are too simple and thus perspective correction is needed to enhance the recognition rate.Due to the complexity of the layout of Chinese name cards text segmentation also attaches strong significance to the performance. In this paper a improved name card recognition system would be devised and implemented. Two new algorithms for perspective correction and text segmentation would be proposed respectively.