In this article, we present an automated solution for capturing facial expressions, directly extracting changes in respiration temperature, and promptly calculating respiration rates. We use the FLIR Lepton 3.5 thermal camera module for temperature detection and the pre-trained facial recognition model SF-TL54 to conduct the remote detection on the Raspberry Pi 4 Model B desktop computer. The proposed system uses the property of temperature decrease during inhalation and increase during exhalation to determine the respiration rate using the temperature variations. The experiment was conducted to adjust trigger points in real time to minimize errors caused by head movements or environmental changes, thereby enabling immediate respiration rate detection. In breathing cessation, the system displays relevant information directly on the screen.
