Removal of border noise, show through and shadow correction in irregularly illuminated scanned document images

Today most of the libraries maintain a digitized version of their document. During the process of scanning the pages of an open book, often problems such as dark borders, show through and shadow are frequently encountered. The shadow is a result of less illumination near the sides of the pages and it degrades the visibility of text near the side. Also any dark areas in the borders mar the appeal of the document. People have carried out various works on removal of such noises in the past. In this paper we have described an effective method based on thresholding to remove show through, combination of horizontal and vertical projection profile and pixel scanning for detection and removal of shadow and dark borders from scanned document images without affecting the useful information contained in the document. The technique was tested on various scanned images and to assess the performance of the method various metrics were calculated and tabulated.