Deep Learning in Document Processing for Enterprise Automation

Authors: Cibaca Khandelwal

DOI: https://doi.org/10.5281/zenodo.14881073

Short DOI: https://doi.org/g84948

Country: USA

Full-text Research PDF File:   View   |   Download

Abstract: Deep learning has significantly advanced document processing, enabling enterprises to automate workflows that were previously labor-intensive and error prone. Document AI systems combine computer vision, natural language processing (NLP), and optical character recognition (OCR) to extract, classify, and interpret information from diverse document types. This paper examines deep learning techniques for document processing, focusing on publicly available datasets such as SROIE, IAM, and FUNSD. An OCR pipeline is implemented and evaluated to highlight the practical implications of these techniques. We discuss challenges, performance metrics, and future research opportunities, providing a roadmap for enhancing document AI’s role in enterprise automation.

Keywords: Deep learning, document processing, enterprise automation, OCR, natural language processing (NLP), computer vision, CRNN, SROIE, FUNSD

Paper Id: 232145

Published On: 2022-02-12

Published In: Volume 10, Issue 1, January-February 2022

Share this