Processing and Isolating Text Information on Business Card

• Main Authors: LIN,TING-SHU, 林廷樹
• Other Authors: SHEU, JIA-SHING
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/5w7f2y

碩士 === 國立臺北教育大學 === 資訊科學系碩士班 === 106 === With technology improving, the functions of mobile device are more complete, and become a part of our life. If mobile device combinate with commercial activity, it will improve the efficiency of business. In this paper, a rapid method of capturing basic information from business cards is investigated. Now the business cards are combined text and illustrations. The text information is more important than the illustrations. In this study, the text area is focus. The area that is not the text is called noise. Removing the redundant and unnecessary noise in order to search target objects easily based on image recognition. The most important parts of data will be gotten which are combined the Canny edge detector, Hough transform and largest quadrilateral segmentation methods. First, reducing the interference from the background of business cards to select the possible text areas from the processed image by both of Canny edge detector and connecting the neighboring text through morphology closing. In addition, the 8-connected components labeling method will confirm the connected text. Finally, in order to locate the accurate position of text areas the length and width of the rectangle will be defined by the largest length and width in same labels. Then, the non-text will be removed by empirical law.