DOCUMENT EXTRACTION USING FORMNET AND TESSERACT OCR

This research focuses on the development of an advanced document extraction system that integrates Tesseract OCR, FormNet, and pdfplumber to efficiently extract text and structured information from both PDF and image files. Traditional OCR systems often struggle with complex document layouts, poor-quality scans, and multi-page processing. To overcome these limitations, the proposed system combines traditional optical character recognition techniques with deep learning capabilities. The inclusion of preprocessing methods such as grayscale conversion and binarization enhances recognition accuracy, while FormNet’s Graph Neural Network (GNN) architecture captures spatial and contextual relationships within documents. The system is designed with a Streamlit-based user interface, allowing non-technical users to upload documents, extract information, and download results in editable formats like DOCX or TXT. The research successfully demonstrates that integrating OCR and deep learning technologies can significantly improve the accuracy, speed, and scalability of document extraction processes. Test cases involving PDF and image files confirmed the system’s ability to extract both structured and unstructured text efficiently, making it applicable in various domains such as healthcare, education, finance, and legal documentation. The user-friendly design ensures accessibility, while the modular architecture supports future integration with cloud and AI-based services. Overall, this study contributes to the field of intelligent document processing by presenting a scalable and reliable solution that minimizes manual effort, enhances data accuracy, and promotes automation in document-centric workflows.