eduzhai > Applied Sciences > Engineering >

Text Extraction from Business Cards and Classification of Extracted Text Into Predefined Classes

  • Save

... pages left unread,continue reading

Document pages: 8 pages

Abstract: Optical Character Recognition (OCR) is the technology for identification of characters with utmost accuracy possible by employing suitable preprocessing, processing and post processing refinements. Practical application of OCR is very wide ranging from day-to-day need to scientific research purposes. One very crucial application is to Automate Digitalisation of Business cards. Since Business cards comes in different fonts and sizes, and most importantly with different lighting conditions, applying OCR can be done after careful processing. Avoiding noise from source image is one of the most crucial step in any image-processing process and it has major weightage in the accuracy of further step and thus indirectly has a huge contribution for our final outcome. Further proper noise cancellation in our source image can reduce number of future steps required to attain good accuracy and also avoid problem of iterating our sample over cycles to avoid better contrast or to distinguish text in our source lying in a noisy matrix. Digitalising business cards aims at classification of the text extracted from our source image in the hard copy of the business card directly into respected following classified fields so that it becomes a lot easier to proceed any desired function ones aims to do with that business card in this online era.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...