Restoration of degraded Kannada handwritten paper inscriptions (Hastaprati) using image enhancement techniques

The digital image processing has got much attention of the researchers towards development of automatic optical character recognition system. The image analysis of documents mainly focused on printed and handwritten document images. In printed document image analysis, the input image is that of machine printed document, where as in handwritten document image analysis, it is of document handwritten by different persons on a paper by using pen or pencil. Presently most of the research work addresses issues related to handwritten document images and the Kannada script. The printed script recognition system is much easier than handwritten script recognition system. In printed documents, font style and size of the characters standardized but handwritten characters vary in size and style of font from person to person and time to time, which is a very tedious job for recognition. In this paper, we proposed a new novel approach for restoration of degraded Kannada handwritten paper inscription (hastaprati) using the combination of special local and global binarization techniques, by eliminating of non-uniformly illuminated background. The performance evaluation is done by calculating the values of MSE and PSNR and these results are compared with manual results obtained by the Epigraphists. It is also compared with other standard methods, namely, souvola and niblack in the literature, which demonstrate the efficacy of the proposed method. Restoration of degraded Kannada handwritten documents plays an important role in age identification, Kannada character recognition and classification for the Kannada handwritten script written on paper (hastaprati).