Mental Stress Detection via Facial Cues

Abstract

Stress has become one of the most common psychological issues in today’s fast-paced lifestyle, adversely affecting both mental and physical health. Traditional methods are either invasive, costly, or subjective. This project proposes a non-invasive, AI- based stress detection system that analyses facial cues in real time. The proposed system offers functionalities such as Login, Signup, Dashboard and history tracking, The system uses OpenCV for video capture and face detection, while dlib’s 68-point facial landmark predictor is applied to extract significant facial features. These features are passed to a Convolutional Neural Network (CNN) model which performs emotion classification.