Removal of Historical Document Degradations Using Conditional GANs , In International Conference on Pattern Recognition Applications and Methods (ICPRAM 2019)

Veeru Dumpala, Sheela Raju Kurupathi, Syed Saqib Bukhari, Andreas Dengel

In: The 8th International Conference on Pattern Recognition Applications and Methods ICPRAM 2019. International Conference on Pattern Recognition Applications and Methods (ICPRAM-2019) February 19-21 Prague Czech Republic Insticc 2019.


One of the most crucial problem in document analysis and OCR pipeline is document binarization. Manytraditional algorithms over the past few decades like Sauvola, Niblack, Otsu etc,. were used for binarizationwhich gave insufficient results for historical texts with degradations. Recently many attempts have been madeto solve binarization using deep learning approaches like Autoencoders, FCNs. However, these models donot generalize well to real world historical document images qualitatively. In this paper, we propose a modelbased on conditional GAN, well known for its high-resolution image synthesis. Here, the proposed model isused for image manipulation task which can remove different degradations in historical documents like stains,bleed-through and non-uniform shadings. The performance of the proposed model outperforms recent state-of-the-art models for document image binarization. We support our claims by benchmarking the proposedmodel on publicly available PHIBC 2012, DIBCO (2009-2017) and Palm Leaf datasets. The main objectiveof this paper is to illuminate the advantages of generative modeling and adversarial training for documentimage binarization in supervised setting which shows good generalization capabilities on different inter/intraclass domain document images

Weitere Links

ICPRAM_2019_59_CR.pdf (pdf, 9 MB )

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence