Learning 2D Morphological Network for Old Document Image Binarization

Document image binarization, especially old handwritten documents, is a very important yet challenging task. There are various bottlenecks for binarizing historical documents due to different types of degradation present imultaneously such as back impression, ink bleed through, faded colours, and wear and tear of the writing media. We consider these degradation as various types of noise in the document image. Here we have proposed a 2D morphological network which consists of basic morphological operation like dilation and erosion to perform our targeted task. The network also includes linear combination of output from dilation and erosion operations. The aforementioned 2D morphological network is applied for image binarization, where the structuring elements (SEs) and the weights of the linear combination layer are learned through back-propagation. The proposed network has been evaluated on DIBCO 2017 and H-DIBCO 2018 and ISI-Letter dataset. Our results show more convincing as compared to the results of other state-of-the-art methods. Though the network is developed for old handwritten documents, it may be tuned to work for image processing task. The source code can be found here https://github.com/ranjanZ/ICDAR_Binarization.

I. Introduction

The world of twenty-first century is entirely shifting towards a paperless world because of today's advancement in technology. In spite of that fact, the importance of the historical documents remained the same. Historical documents may be a family letter, a document of national importance or old machine printed valuable book. All these types of documents describe the socioeconomic condition of that period. Besides those information, the handwritten documents may also describe the personality and mood of the writer. For extracting the information from the historical documents, binarization is an important pre-processing step. But the aging of the pages, bleeding of ink, wear and tear of the writing medium and fading of the inks cause deterioration and noise in the binarized image.