A New and Efficient Technique to Remove Back-to-Front Interference in Historical Document Images

Abstract

The study of historical records is a topic that presents significant difficulties for specialists from different fields, for example, history, political science, brain research, software engineering, among others. Historical archives contain signifi- cant information about cultural and scientific value. Historical artifacts comprise of archives, letters, daily papers, pictures, maps, and so on. A large number of these are put away in libraries, historical centers or government files. In any case, because of the conservation, few individuals have admittance to this material. Also such documents are frequently degraded over time. The main aim of preserving historical archives is to reestablish the corrupted content containing data. Digital preservation is the foremost requirement of today to preserve the historical documents. Digital preservation is the most promising method for preserving the historical documents. But if the document is old age document then this method may not give the best results and may give results which might be dif- ficult to read. Now this further enhances the problem if the document is written on both sides, there may be some impression of back image to the front image, which reduces the readability of the document. This phenomenon is known as “back-tofront interference” also called “bleeding” or “show-through”. The thresholding of such records with conventionally used algorithm produces unreadable documents. Generally numerous binarization strategies are implemented in the literature for various sorts of binarization issues. The few simple available binarization strategies cannot be applied to numerous binarization issues. Keeping in mind the end goal to enhance the quality of historical archives pictures, a joined methodology based on mathematical morphology, global and local binarization strategies is applied. This technique at initial step removes the background (noise) and adjusts the intensity values of pixels of the historical image, which enhances the performance of global thresholding. After that, the mathematical erosion operation is applied on this binarize image, then local binarization is applied to binarize the image. The proposed technique effectively denoises different sorts of corrupted records, enhancing textures with clear background. The present study proposed new and efficient technique to eliminate back-to-front type of interference in such type of historical artifacts