Skip navigation
Please use this identifier to cite or link to this item:
Title: Improving Document Binarization Via Adversarial Noise-Texture Augmentation
Authors: Bhunia A.K.
Bhunia A.K.
Sain A.
Pratim Roy, Partha
Published in: Proceedings of International Conference on Image Processing, ICIP
Abstract: Binarization of degraded document images is an elementary step in most problems involving document image analysis. The paper re-visits the binarization problem by introducing an adversarial learning approach. We construct a Texture Augmentation Network that transfers the texture element of a degraded reference document image to a clean binary image. In this way, the network creates multiple versions of the same textual content with various noisy textures, thus enlarging the available document binarization datasets. Finally, the newly generated images are passed through a Binarization network to get back the clean version. By jointly training the two networks we can increase the adversarial robustness of our system. The most significant contribution of our framework is that it does not require any paired data unlike other Deep Learning-based methods [1], [2], [3]. Such a novel approach has never been implemented earlier thus making it the very first of its kind in Document Image Analysis community. Experimental results suggest that the proposed method1 achieves superior performance over widely used DIBCO datasets. © 2019 IEEE.
Citation: Proceedings of International Conference on Image Processing, ICIP, (2019), 2721- 2725
Issue Date: 2019
Publisher: IEEE Computer Society
Keywords: Adversarial Learning
Document image binarization
Style transfer
Unpaired data
ISBN: 9.78154E+12
ISSN: 15224880
Author Scopus IDs: 57188719920
Author Affiliations: Bhunia, A.K., Nanyang Technological University, Singapore
Bhunia, A.K., Jadavpur University, India
Sain, A., Cognizant Technology Solutions, India
Roy, P.P., Indian Institute of Technology, Roorkee, India
Corresponding Author: Bhunia, A.K.; Nanyang Technological UniversitySingapore; email:
Appears in Collections:Conference Publications [CS]

Files in This Item:
There are no files associated with this item.
Show full item record

Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.