Binarization techniques for degraded document images — A review

Abstract

Document Image binarization is the segmentation of the document into foreground text and background. It is done to obtain the clear images from which text can be retrieved easily. Thresholding is used for the segmentation of the document images. This paper, presents a review on various document image binarization techniques. Evaluation performance metrics used for the evaluation of the binarization techniques are also explained. Comparison of the performance of the binarization techniques based on the performance metrics like PSNR, F-Measure, NRM and MPM is shown. Performance of the techniques is evaluated on the dataset of DIBCO-2009 and DIBCO-2010.