Enhancement of ancient manuscript images by log based binarization technique

Enhancement of ancient manuscripts is a predominant field in the world of image processing. Binarization is the essential tool to recognize the text through OCR technologies for the preservation of the ancient manuscripts. This paper addresses the problem of enhancing and binarizing the deteriorated old manuscripts by a simple novel method called Log based binarization. The proposed method modifies the pixel intensity so as to better differentiate the black pixels and thus produce better binarization results, has been explored. Detailed comparison of existing thresholding algorithms with the new algorithm has been given. The inspiration for the new algorithm was drawn from the standard Niblack algorithm and Nick Algorithm.