Sep 9–12, 2019

Improving OCR quality of documents using generative adversarial networks

11:55am12:35pm Wednesday, September 11, 2019
Location: 230 C
Average rating: *****
(5.00, 1 rating)

Who is this presentation for?

  • Directors and VPs of operations involving document processing




While we move toward a digital world, paper-based document processing is still an integral part of business processes across a wide range of domains from finance to healthcare. Traditionally, these documents are processed manually and natural language processing (NLP)-based document processing solutions bring significant efficiencies in this manual process by automating document classification, extraction, or search. OCR is a crucial step in any document-processing task as it enables machine interpretation of text and parsing textual information. However, issues in document quality, such as blurred text, fragmented characters, merging characters, low resolution, skew, and noise, can tremendously affect OCR accuracy and degrade its performance. In particular, images are corrupted through stains, wrinkles, pixel noise, and impulse noise added during scanning. Poor resolution leads to the merging of character strokes with the document background.

Nagendra Shishodia, Solmaz Torabi, and Chaithanya Manda explore the general adversarial network (GAN) developed at EXL to overcome these challenges of enhancing the resolution and denoising scanned images. The model was trained using custom generated data combined with externally available datasets. Once trained, the generator’s able to increase the resolution of the document by a predefined factor, sharpen the character borders, increase contrast, and eliminate pixel noise while preserving edge features of characters. The model significantly improves OCR accuracy (using open source and commerical OCR systems) on a word and character level. Implementing the GAN model also helped the company achieve significant efficiencies in multiple document-processing solutions from finance to healthcare. The model also helped improve accuracy of handwriting detection APIs.

You’ll learn how GANs can bring efficiency and accuracy in document-processing pipelines. In particular, Nagendra, Solmaz, and Chaithanya detail how the dataset was created for this task, the model architecture and training methodology, and they showcase the performance of the model across multiple document-processing applications and examine potential next steps.

Prerequisite knowledge

  • General knowledge of deep learning and GAN

What you'll learn

  • Understand how GANs improve document quality and OCR accuracy
Photo of Nagendra Shishodia

Nagendra Shishodia


Nagendra Shishodia is the head of analytics products for EXL, where he leads the analytics product development initiative and has written thought leadership articles on healthcare clinical solutions and AI. He has over 17 years of experience in developing advanced analytics solutions across business functions. His focus has been on developing solutions that enable better decision making through the use of machine learning, natural language processing, and big data technologies. Nagendra consults with senior executives of global firms across industries including healthcare, insurance, banking, retail, and travel. Nagendra holds an MS degree from Purdue University and a BTech from the Indian Institute of Technology Bombay.

Photo of Solmaz Torabi

Solmaz Torabi


Solmaz Torabi is a data scientist at EXL, where she’s responsible for building image and text analytics models using deep learning methods to extract information from images and documents. She holds a PhD in electrical and computer engineering from Drexel University.

Photo of Chaithanya Manda

Chaithanya Manda


Chaithanya Manda is an assistant vice president at EXL, where he’s responsible for building AI-enabled solutions that can bring efficiencies across various business processes. He has over 10 years of experience in developing advanced analytics solutions across multiple business domains. He holds a bachelor’s of technology degree from the Indian Institute of Technology Guwahati.

Comments on this page are now closed.


Anirudh Kalyanaraman |
07/21/2019 10:03pm PDT

I am a budding Data Scientist…Need initial input for me to clear the noise from an Image for me to try some interesting

  • Intel AI
  • O'Reilly
  • Amazon Web Services
  • IBM Watson
  • Dataiku
  • Dell Technologies
  • Intuit
  • Gamalon
  • Hewlett Packard Enterprise
  • MapR Technologies
  • Sisu Data
  • Intuit

Contact us

For conference registration information and customer service

For more information on community discounts and trade opportunities with O’Reilly conferences

Become a sponsor

For information on exhibiting or sponsoring a conference

For media/analyst press inquires