Stress Detection Based on Facial Expressions using a Combination of CNN and KNN

This study describes a method for facial expression-based stress detection that uses Convolutional Neural Network (CNN) and K-Nearest Neighbors (KNN) to assess and detect stress based on facial expressions. Accurate identification of stress is essential because it has a major impact on both personal health and productivity. In particular, MediaPipe and the Facial Action Coding System (FACS) are used for feature extraction and stress level assessment in the suggested method's integration of AI and ML technologies. KNN and CNN models work together, obtaining a 95% accuracy rate in stress detection. A collection of 7,066 face photos representing seven different emotional states is used to assess the system's performance. The combined model performs better than the separate KNN and CNN models, according to the experimental data, indicating its potential for use in psychological testing and health monitoring.