Use of adaptive methods to improve degraded document images

Abstract

An adaptive method for enhancing and binarizing the images of degraded documents has been presented here. This method does not need any feature handling by the user and it handles all kinds of degradations, removes noise, ensures connectivity of stroke and improves low-contrast. The project briefly includes following steps: a pre-processing procedure using a low-pass Wiener filter to produce a smoothened image, an approximate estimation of foreground regions,a background surface calculation by interpolating neighboring background intensities, a thresholding by combining the calculated background surface with the original image followed by a post-processing step which is carried out to enhance the quality of foreground regions and preserve line connectivity of texts.