A Hybrid OCR-CNN-Metadata Model for Academic Documents Authentication

Document forgery has become a serious concern in digital services such as banking, education, recruitment, and government verification systems. Manual verification is time-consuming, error-prone, and not scalable. This research proposes an AI-based document verification system that combines Optical Character Recognition (OCR), Convolutional Neural Networks (CNN), and metadata analysis to verify the authenticity of digital documents. The system performs image forgery detection, text consistency verification, and metadata anomaly checking to generate a final authenticity score. By integrating visual, textual, and hidden metadata features, the proposed approach improves reliability, reduces false verification, and supports automated decision-making. Experimental analysis demonstrates that the hybrid model outperforms traditional single-technique verification methods and is suitable for real-world document authentication systems.