“…The previous approach by Moghaddam and Cheriet [7] was based on an adaptive multi-scale binarization algorithm. Bhardwaj et al [8] presented a new technique for photo defect detection, by sensing the light after it passes through the photo and then thresholding the digitized image to identify defect objects. Zhang et al [9] proposed a framework, based on an inpainting algorithm, and a shape-from-shading technique, to remove unwanted distortions in digitized image from printed documents.…”