Intellectual Property India Publishes Patent Application for An Intelligent Automated Document Processing System With Adaptive Multi-Stage Recognition And Context-Aware Error Correction, Based On Deep Learning Optical Character Recognition

MUMBAI, India, Aug. 24 -- Intellectual Property India has published a patent application (202641097940 A) filed by Sri Eshwar College Of Engineering on August 12, 2026, for An Intelligent Automated Document Processing System With Adaptive Multi-Stage Recognition And Context-Aware Error Correction, Based On Deep Learning Optical Character Recognition.

Inventors include Dr. M. P. Geetha; and Ms. C. Saranya.

The application for the patent was published on August 21, 2026, under issue no. 34/2026.

Abstract: According to this disclosure, an Optical Character Recognition (OCR) system by deep learning for automatically document processing is provided, and it comprising: a document acquisition module; an adaptive image preprocessing module configured to improve an image quality of a document image; a document layout analysis module configured to analyze a structure of the document; a deep learning based text recognition module configured to textual content retrieved from the document in image form; a contextual error correction module configured to improve recognition accuracy of the textual content retrieved from the document in image form; and a structured information extraction module configured to convert the text extracted from the document image into a structured form. The preprocessing module performs image preprocessing on the acquired document image and improve the quality of the text recognition, including noise filtering, image enhancement, skew correction and document segmentation. The deep learningbased text recognition module is based on a hybrid deep learning architecture capable of extracting both printed and handwritten text from structured and unstructured documents. The contextual error correction module uses a language-aware semantic analysis and recognition to achieve high recognition accuracy in a variety of multilingual documents. The system converts information extracted from documents to a structured digital data form, integrates with document management, cloud storage, and workflow automation software to support real-time, batch document processing, and is applicable in fields of healthcare, banking, education, legal affairs and government administration.

Disclaimer: Curated by HT Syndication.