<![CDATA[The advancement of digital technology and artificial intelligence has opened vast opportunities for intelligent health monitoring systems that operate automatically, in real time, and without physical contact. This study aims to develop a system for detecting human body temperature and respiratory patterns using an infrared thermal camera based on digital image processing and machine learning. The research method involves thermal data acquisition on facial areas (forehead, nose, and mouth), image preprocessing using two-point temperature calibration and Gaussian filtering for noise reduction, and segmentation of the respiratory region using the adaptive thresholding method. Feature extraction is performed by analyzing temperature variations in the nose and mouth regions as thermal signals, which are converted into the frequency domain using the Fast Fourier Transform (FFT) algorithm to determine the respiration rate. Classification is carried out using the Support Vector Machine (SVM) algorithm to distinguish three physiological conditions: normal, fever, and respiratory disorder. The dataset consists of 550 thermal images, divided into 385 images (70%) for training and 165 images (30%) for testing. Experimental results show that the system achieves an accuracy of 98.32%, with an estimated forehead temperature of 145.23°C (a relative value from initial calibration) and a respiration rate of 6.6 bpm, indicating the subject’s condition as fever. This study demonstrates that the combination of thermal image processing, FFT algorithms, and SVM classification is effective for non-invasive, high-precision, and efficient health monitoring systems. The proposed system has the potential to support the development of the Internet of Medical Things (IoMT) for safe, accurate, and adaptive remote health monitoring in response to patients’ physiological changes]]>
Copyrights © 2025
