USING HEURISTIC ALGORITHMS ON THERMAL IMAGES FOR KNEE OSTEOARTHRITIS

Abstract

In individuals with osteoarthritis (OA), the temperature in the knee area where osteoarthritis occurs is higher than in healthy individuals. In this study, it is aimed to develop a method that can be used in the early detection of osteoarthritis by processing the images obtained from the thermal camera by using this feature of osteoarthritis. For this purpose, Support Vector Machines and VGG-16 architecture are used as the methods in this study. In the study, thermal images were taken from 998 different individuals using the FLIR E45 thermal camera and processed using Support Vector Machines and VGG-16 architecture. 284 of these thermal images were obtained from sick individuals and 714 thermal images were obtained from healthy individuals. Visual inspection and image resizing pretreatment tasks are performed for thermal images. 224x224 input image size is used for VGG-16. Deep learning algorithms and libraries were used in the study. Infrared thermography (IRT) reveals the related disease by emphasizing the asymmetric behavior that occurs in thermal color maps in both knees. The results obtained in the study clearly show that temperature can be considered as a key parameter in the assessment of discomfort.