Document Image Binarization

Documents often get affected by noise due to various issues like casual handling occurred during storing, or image acquisition processes. These noises not only affect the quality of the documents but also degrade the visual appearance of the content. Therefore, these have high potential to downgrade the final outcome. To suppress the noise, occur in an input document image at the initial stage, it is essential to use an efficient pre-processing before further analysis. Pre-processing converts the input image to a specific form, suitable for further analysis without disturbing the knowledge content. Among all, binarization is very popular and commonly performed pre-processing before many document image processing tasks. Therefore, in the present chapter, a detailed discussion on document image binarization and the recent progresses in this domain are presented. Additionally, in this chapter, various degradations related to document image, and various noise models are also described to make the discussion complete.