DeepShield – Real-Time Deepfake Detection for Social Media
AI-powered system that detects deepfake videos by analyzing facial inconsistencies, lip-sync mismatches, and frame anomalies to verify video authenticity.
TruthLens is an AI-based deepfake detection platform designed to help users identify manipulated videos and prevent the spread of misinformation on the internet.
With the rapid advancement of generative AI technologies, deepfake videos are becoming increasingly realistic and difficult to detect with the human eye. These videos are often used to spread fake news, impersonate individuals, manipulate public opinion, and conduct online scams. Currently, most available detection tools are either limited to researchers or require technical expertise, making it difficult for everyday users to verify the authenticity of online videos.
This project proposes a user-friendly system that allows users to upload or paste a video link, after which the system analyzes the video using deep learning techniques. The model examines facial expressions, lip synchronization, blinking patterns, and frame-level inconsistencies to determine whether the video is genuine or AI-generated.
The system then provides a deepfake probability score along with visual explanations highlighting suspicious frames. By making deepfake detection accessible to the public, the platform aims to strengthen digital trust, reduce misinformation, and promote responsible AI usage.
The system can be integrated with social media platforms, news verification tools, and browser extensions to provide real-time video authenticity checks.