Detection of Deepfake Images and Videos Using SVM, CNN, and Hybrid Approaches
Ivan Stefanov Stankov, Evgeni Evgeniev Dulgerov · 2024
The proliferation of deepfake technology, driven by advancements in artificial intelligence, particularly generative adversarial networks (GANs), has introduced new challenges in verifying the authenticity of multimedia content. Deepfakes can create highly realistic fake images and videos that are difficult to distinguish from real ones, posing significant threats to privacy, security, and information integrity. This study investigates the effectiveness of three machine learning models—Support Vector Machine (SVM), Convolutional Neural Network (CNN), and a combined CNN-SVM approach—in identifying deepfake images and videos. By leveraging a dataset containing both real and computer-generated content, we trained and evaluated each model on their ability to accurately classify these media types.