Deepfake Face Detection Using Machine Learning
Deepfake Face Detection Using Machine Learning
Authors:
Asmita Sudam Gaikwad1, Swati S. Hinge2
1Department of Artificial Intelligence, Sanghavi college of Engineering, Nashik
2HOD Artificial Intelligence, Sanghavi college of Engineering, Nashik
Abstract—Deepfake technology has become one of the most rapidly growing fields in artificial intelligence and computer vision. While deepfake generation systems are capable of creating highly realistic manipulated videos, they also introduce serious threats related to misinformation, digital fraud, political manipulation, and cybercrime. This research paper presents a deep learning based deepfake face detection system that combines ResNext Convolutional Neural Networks and Long Short-Term Memory networks for accurate classification of manipulated and authentic videos. The proposed system extracts spatial features from individual frames and temporal features from frame sequences to achieve improved detection performance. The system is implemented using Django, PyTorch, OpenCV, and CUDA acceleration for efficient real-time processing. Experimental evaluation demonstrates high detection accuracy of approximately 97% using 100-frame sequence models. The paper discusses the architecture, dataset preparation, preprocessing pipeline, implementation methodology, experimental results, limitations, and future enhancements of the proposed system.
Key Words: Deep-Fake face Detection, Real-Time Video Analysis, Convolutional Neural Networks, Gated Recurrent Unit, Artificial Intelligence, Video Manipulation Detection