C069

DEEP LEARNING-BASED IMAGE FORGERY DETECTION

MR. TAY HAN CHUNG, MS. PA PA MIN

AFFILIATION
Faculty of Information Science & Technology, Multimedia University

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.