Hybrid Deep Learning Framework for Real-Time Stress Recognition Using Infrared Thermal Facial Imaging

Stress is a major contributor to physical and psychological disorders, making its timely detection crucial in healthcare, human–computer interaction, and occupational safety. Conventional stress assessment techniques rely on contact-based physiological sensors, which are intrusive and unsuitable for continuous monitoring. Infrared thermal facial imaging provides a non-contact alternative by capturing temperature variations related to autonomic nervous system activity. This study proposes a hybrid deep learning framework combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks for real-time stress recognition. The CNN extracts discriminative spatial thermal features from facial regions, while the LSTM models temporal thermal dynamics for continuous stress monitoring. Experimental evaluation on a thermal stress dataset demonstrates superior performance compared to single-model approaches, achieving high accuracy and robustness under real-time constraints.