Downloads: 0
Research Paper | Computer Science and Information Technology | Volume 15 Issue 4, April 2026 | Pages: 1419 - 1423 | India
Enhanced Image and Document Forgery Detection Using Artificial Intelligence
Abstract: The rapid advancement of digital editing tools and generative artificial intelligence technologies has rendered sophisticated content manipulation increasingly accessible, posing an unprecedented threat to the integrity of visual and documentary information across critical domains. This paper presents a high-precision, automated framework for enhanced image and document forgery detection employing state-of-the-art Artificial Intelligence and Machine Learning architectures. The proposed system leverages transfer learning with pre-trained deep convolutional neural networks, specifically VGG16 and ResNet50, to perform granular feature extraction and identify subtle structural anomalies that are imperceptible to human observation. For digital imagery, the model detects complex spatial inconsistencies arising from copy-move operations, image splicing, and inpainting techniques. For document forensics, the system authenticates sensitive records by identifying forged signatures, fraudulent seals, and digitally substituted textual content. The system achieves an overall classification accuracy of 95%, with a Precision of 93%, Recall of 94%, and an F1-Score in the range of 93?94%. A Flask-based web application delivers real-time predictions accompanied by Grad-CAM heatmaps for result explainability and transparent forensic reporting. The proposed framework provides a scalable and dependable security solution for government agencies, legal institutions, and online media platforms requiring reliable digital content authentication.
Keywords: Image Forgery Detection, Document Forensics, Transfer Learning, VGG16, ResNet50, Support Vector Machine, Grad-CAM, Deep Learning, CNN, Flask
How to Cite?: Bibin K Shaji, Bindu B, "Enhanced Image and Document Forgery Detection Using Artificial Intelligence", Volume 15 Issue 4, April 2026, International Journal of Science and Research (IJSR), Pages: 1419-1423, https://www.ijsr.net/getabstract.php?paperid=SR26422152329, DOI: https://dx.dx.doi.org/10.21275/SR26422152329