Recognition And Information Extraction In Camera Captured Business Card Image

Business card, as a concise and carrier, has played a very important role in today's commerce, in which there is much useful information. Business cards have brought convenience to people, but it is boring to manage them. Some mobile phones and PDAs are able to manage business cards, but less have ability to implement automatic input or the effect is not ideal. This paper focuses on English business card image segmentation and recognition and presents corresponding algorithms. We first introduce the main system framework of English business card recognition, including image preprocessing, document image analysis, feature extraction, character recognition processes, content information understanding. Then introduce and describe every function module and process of main framework and compare with relevant algorithms, especially text areas extraction and content understanding.In this paper, an algorithm is presented for detecting text areas in business card from color image captured by mobile phone camera. It is carried out by three steps: edge detection using sobel arithmetic operators in four directions, connected component analysis, text area candidate detection and decision. Experimental results demonstrate that the proposed approach has good performance for segmenting text areas in mobile phone camera captured images with varying font-sizes, font-colors, background and languages. And then, according to the corresponding area segmentation, document image analysis and character recognition processes were made. At last, after analyzing text area and finding address area and name area, the contact information (like name, telephone, fax, e-mail) on the business card are picked up and stored into database. The experiment shows that the proposed method validates business card recognition system.