Human Emotion Detection Using Thermal Images

In healthcare, security, and even in human-computer interactions, the application of emotion detection technology is becoming increasingly popular.  Exiting methods employing light pictures have problems with occlusion and illumination.  This project uses thermal imaging to identify the changes in facial temperatures due to emotions, which guarantees functionality in all lighting conditions.  It also uses real-time emotion detection from thermal face data using an optimized YOLOv5 model. The model is trained using labeled thermal datasets and is modified using transfer learning to streamline the model for specific features of the domain.  This approach ensures emotion detection is done fast, accurately, and stealthily.  This is critical in low-light situations or where privacy is paramount.  Examples of such include access control, adaptive learning, and mental health care.  The technology enhances emotional computing by integrating deep learning with infrared imaging.