Review on Deep Learning-Based Intelligent Document Forgery Detection and Secure Verification Framework

Preksha, Divya G S, S Sivaramasubramanian M R, Umashanker L

In recent years, there has been a dramatic expansion of digital communication, online document sharing, and therefore the importance of a highly secure document verification system has risen greatly. Existing systems require manual effort, slow verification process and are susceptible to forgery. In this paper, we develop an AI based secure document verification and monitoring framework using the combination of different deep learning approaches which results in the high accuracy detection of authenticity and the efficient implementation of real-time monitoring and automated verification of documents. The framework proposed in this paper utilizes the benefits of Optical Character Recognition (OCR), Convolutional Neural Networks (CNN), face matching algorithm, validation of QR code, and a secure storage mechanism based on blockchain technology to secure document from unauthorized changes. It also supports detection of signature forgery, fake images and fake certificate in documents, false authentication in educational institutions etc. Additionally, it has an IoT based real-time monitoring and alerting system utilizing the cloud connectivity. We provide evidence that through the conducted experimental analysis, our approach not only significantly increases the verification accuracy and decreases the manual effort required but also proves to be less susceptible to fraud and enhances the security against conventional system. This can be applied in various applications such as educational institutions, banking systems, and governmental sectors, healthcare records, etc. And also, for digital identities verification platforms.