An approach for improving Optical Character Recognition using Contrast enhancement technique

Abstract Digital cameras and mobile image capture of documents are two examples new developments in the fields optical character recognition text recognition. Scans or photographic images even natural photography results can be distorted to point where OCR digitization is inaccurate. It offers a unique non-parametric unattended approach correct unwanted document distortions achieve optimal accuracy. applies highly effective stack enhancement algorithms restore perfect by unknown sources distortion. First, it provides means modifying local brightness contrast order better handle different illumination levels atypical light transmission patterns image. Then apply nifty grayscale conversion method your photo give look. Third, uses unsharp masking techniques further enhance important details images. Finally, we use best global binarization technique prepare final for The proposed has potential significantly improve rate accuracy