Optical character extraction under different illumination conditions application in real time systems
Avani Oza, Paras Gosai · 2016
This paper presents algorithm for character extraction of paper printed characters using MATLAB computation software with image processing toolbox. The printed paper is digitized using digital camera. Inconsistent lighting conditions while digitizing the paper severely affects the result. The main focus of this paper is to extract the characters in various illumination conditions. The planned method is an adaptive image thresholding method which carries separation of the characters on the front from its backdrop. Adaptive thresholding method chooses the finest threshold value from single pixel to maximize the inter class variance of the black and white pixels. This paper introduces chronological manner for scanning the text for evaluation of single line in real time. This means to be able to extract words facing the pointed finger as it moves across the text.