Emotion-Aware Safety System for Industrial Environments

This project develops an emotion perception system for factory workers aimed at enhancing workplace safety through intelligent human-state monitoring. Using Google MediaPipe for facial landmark extraction and a hybrid rule-based + LSTM model, the system identifies and predicts workers’ emotional states in real time from live video streams.
By combining facial expression features, temporal modeling, and historical risk data, the system detects early signs of fatigue, stress, or frustration—allowing preemptive intervention before safety-critical incidents occur. The framework provides a foundation for emotion-aware industrial AI, bridging affective computing and occupational safety research.
Built with Python, TensorFlow, and OpenCV, this project demonstrates how multimodal sensing and temporal learning can help build empathetic, human-centered factories where technology safeguards both productivity and well-being.