Fake Document Detection

Merin Ann George, Shyma Kareem

Abstract: The growing worry about the accuracy of computerized and manual reports has raised the need of professional and reliable methods for the detection of forged or altered documents. This particular project aims to develop an intelligent Fake Report Detection Framework that employs Artificial Intelligence (AI), Machine Learning (ML), and Digital Forensics to automatically detect and flag false documents. With the help of image processing techniques and advanced anomaly detection models, the system analyzes documents for discrepancies in textual and visual elements, including signatures and logos. Along with text retrieval, the framework's key functionalities include visual component examination, peculiarity detection using machine learning algorithms. It also issues a detailed report stressing areas of concern together with a confidence score assessing the document's authenticity. A web application developed using the Django framework allows users to conveniently upload documents and receive instantaneous evaluations. The project utilizes several datasets of authentic and synthetic documents to train machine learning models, incorporating preprocessing techniques of document image and signature verification. The expected results will significantly increase the reliability of document authentication.

Keywords: Artificial Intelligence, Machine Learning, Digital Forensics