Deepfake Video Detection Using CNN and Flask
MOONA NIVEDITHADEV., CHINTHAGINGALA VASUNDHARA · International Scientific Journal of Engineering and Management · 2025
In recent years, the rise of deepfake technology has posed a significant threat to the authenticity of digital media, enabling the creation of highly realistic yet manipulated videos that can deceive, misinform, and harm reputations. This project proposes a deepfake video detection system that utilizes Convolutional Neural Networks (CNN) for accurate frame-level classification, integrated within a Flask-based web application for user-friendly interaction. The system accepts video uploads via a web interface, extracts key frames using OpenCV, and classifies them as real or fake using a pre-trained CNN model trained on deepfake datasets. The final verdict is generated through aggregated predictions across frames, providing users with a reliable assessment of video authenticity. This solution offers an efficient, scalable, and accessible tool for combating deepfake content and enhancing digital media forensics. INDEX TERMS: Deepfake detection, CNN, Flask, Video forensics, OpenCV, Fake video identification, Digital media authentication, Deep learning.