Digital Business Card Recognition System Based On IOS

In recent years, with the rapid development of the country, Traditionalmanagement of business card is becoming more and more cumbersome. But withthe development of embedded technology, the performance of mobile phones hasalso experienced an explosive improvement. Mobile phones are capable enough tostore and manage the information of business cards. In addition, the versatility ofthe mobile phone has greatly simplified the demands for image recognition.In this project, iPhone is the hardware environment which is based on iOS(iPhone Operating System) system and programed with Objective-C programminglanguage. This project is focus on recognizing the contact information on a businesscard image by iPhone device, generating a contact based on the recognized result,and adding the contact to iPhone’s address book. The business card image can becaptured by iPhone’s camera or downloaded through network.This system can be divided into four subsystems: image acquisition subsystem,image processing subsystem, data processing subsystem, data transmittingsubsystem. Image acquisition subsystem can get a picture from iPhone’s camera oralbum. The image processing subsystem is carried out by two steps: imageprocessing and OCR (Optical Character Recognition). The image processing isbased on an open source library: OpenCV. According to the characteristic ofbusiness cards, some special methods are created to process image. The image isprocessed by four steps: boundary locating and extracting can eliminate theinference caused by background, four directions sobel edge detection can locate allcharacters and pictures of business card, adaptive threshold can change the color ofthe picture to black and white, and connected text area recognizing and extracting isused to locate and extract character areas for recognizing. The optical characterrecognition is based on another open source OCR engine: Tesseract OCR engine.The data processing subsystem can match the result of OCR with the regularexpressions to recognize the particular information type of contact. At last, all theinformation of contact can be transmitted and saved to iPhone’s address book bydata transmitting subsystem.With the overall system testing, this system can recognize all the informationon business card and classify phone number, email address, and URL address. There is no memory leak during running the system. Based on testing results of20business cards with different backgrounds and formats, the average recognitionaccurate and speed of this system can be an acceptable result.