Description of Invention
This project applies deep learning—specifically, the VGG19 pre-trained model—to classify image forgeries into authentic, copy-move, and splicing categories. By combining two benchmark datasets (CoMoFoD and CASIA 2.0) and applying data augmentation, the model is trained to robustly detect various forgery types, ensuring improved image authenticity verification.