Discrete orthogonal moments based framework for assessing blurriness of camera captured document images

Abstract

One of the most widely used tasks in the area of image processing is automated processing of documents, which is done using Optical Character Readers (OCR) from document images. The most common form of distortion in document images is blur which can be caused by defocus, motion, camera shake etc. In this paper we propose a no reference image sharpness measure framework using discrete orthogonal moments and image gradients for assessing quality of document images and validated the results against state of the art image sharpness measures and accuracy of three well known Optical Character Readers.