Web-Based Document Classification System Using Deep Learning

Abstract

Managing large volumes of digital documents is often time-consuming . This project presents a Web-Based Document Classification System that automates the sorting of digital documents. With secure authentication and bulk upload support, users can easily manage files through a drag-and-drop interface. A ResNet18 deep learning model extracts features and classifies documents into categories such as Email, Resume, and Scientific Publication. Results with confidence scores are displayed in an interactive dashboard, offering search, filter, and export options. The system enhances efficiency in corporate, academic, and HR domains by delivering a secure, scalable, and userfriendly solution. Furthermore, the system is designed with modular architecture, allowing easy integration of additional document categories and continuous model improvement for enhanced accuracy over time.