Ghost Protocol – Genuine Face & Emotion Detection System
Ghost Protocol is an advanced computer vision project designed to push the boundaries of accurate face detection. Unlike standard detection systems, this application focuses on authenticity, specifically engineered to distinguish between genuine and posed expressions. By leveraging a multi-stage AI pipeline, it offers real-time analysis of micro-expressions and smile authenticity.
🚀 Tech Stack & Pipeline
The core logic follows a sophisticated sequential pipeline: MediaPipe → FACS Engine → Duchenne Detection → Mistral AI.
- Google MediaPipe: Utilized for high-fidelity facial landmark tracking and keypoint extraction.
- FACS Engine: Implements the Facial Action Coding System logic to break down facial movements into specific action units for emotion derivation.
- Duchenne Algorithm: A specialized algorithm used to validate smiles by analyzing the orbicularis oculi muscle movement (detecting "smiling with the eyes" vs. fake smiles).
- Mistral AI: Employed for high-level semantic analysis to interpret complex emotions and micro-expressions, adding a layer of nuanced understanding to the raw data.
💡 Key Features
- Real-time Landmark Tracking: Instant visualization of facial mesh and key points.
- Authenticity Verification: specifically targets the "Duchenne marker" to validate the genuineness of happiness/smiles.
- Semantic Analysis: Goes beyond simple labeling by using LLMs (Mistral) to contextualize emotional states.
- Web Accessibility: Fully deployed web interface for easy testing and demonstration.
🔗 Project Links:
- Live Application: ghost-protocol-detect.netlify.app
- Source Code: GitHub Repository
Ghost Protocol serves as a robust proof-of-concept for developers and researchers interested in the intersection of biometric analysis, psychology, and artificial intelligence.
