MUMBAI, India, June 22 -- Intellectual Property India has published a patent application (202621048921 A) filed by Moresh Madhukar Mukhedkar; Aditya Singh Bhadauria; Anshul Animesh Mahajan; Shrirang Rajendra Bondre; Sumant Prakash Tiwari; Shakil Tamboli; and Shweta Koparde on April 17, 2026, for Intelligent Invoice Processing: An Ocr & Deep Learning Approach.
Inventors include Moresh Madhukar Mukhedkar; Aditya Singh Bhadauria; Anshul Animesh Mahajan; Shrirang Rajendra Bondre; Sumant Prakash Tiwari; Shakil Tamboli; and Shweta Koparde.
The application for the patent was published on June 12, 2026, under issue no. 24/2026.
Abstract: The Multilingual Invoice Extraction System Using AI is an advanced document processing solution designed to automate the extraction of structured data from invoices written in multiple languages and formats. The system integrates Optical Character Recognition (OCR), language detection, translation mechanisms, and Large Language Models (LLMs) to accurately interpret and process invoice content. It supports various document types including scanned images and digital PDFs, enabling flexible and scalable data handling. The system performs preprocessing techniques such as noise reduction and normalization to improve text extraction quality. The extracted text is further analyzed using NLP and LLM-based models to identify and extract key invoice entities such as invoice number, date, vendor details, tax information, and total amount. The extracted data is then converted into structured formats like JSON or database records for seamless integration with business systems. By combining multilingual understanding with intelligent data extraction, the system reduces manual effort, minimizes errors, and enhances efficiency in invoice processing. It is particularly useful for organizations dealing with large volumes of diverse and multilingual documents, providing accurate, scalable, and automated solutions for real-world applications.
Disclaimer: Curated by HT Syndication.
