Abstract:
Deepfake videos pose a significant challenge in today’s digital landscape, where misinformation and manipulation thrive. In this study, we focus on improving the accuracy of deepfake video detection methods. We explore three techniques: CurveletQA, SSEQ, and Friquee, aiming to enhance the efcacy of existing detection systems. Our experimental results reveal that Friquee consistently outperforms the other techniques in terms of accuracy. By leveraging advanced feature extraction mechanisms and machine learning algorithms, Friquee demonstrates superior capabilities in distinguishing between authentic and manipulated video content. These fndings underscore the importance of continually refning detection methodologies to combat the proliferation of deepfake videos in online platforms.
