A case study: improving face detection in Cartagena vessel traffic service operators to detect fatigue onset using thermographic images through the YOLO5Face model

Fatigue is a common physiological state that can have a significant impact on human performance and well-being, so its early detection is essential in a number of fields, such as maritime safety, which is the subject of this article. Traditional methods of fatigue assessment methods often rely on subjective self-reports or performance metrics, which can lack sensitivity and objectivity. In recent years, thermography has emerged as a promising non-invasive tool that has been validated to detect the onset of fatigue in humans. One way to detect the onset of fatigue is to monitor the temperature of the orofacial area using thermographic imaging. This article provides an overview of the principles of thermography, focusing on the analysis of orofacial thermographic images in real work situations in 16 port control operators in Cartagena. For this purpose, images of these subjects were taken during their watch and the points of interest mentioned above were identified using the YOLO5Face model, a version of YOLO5 trained to recognise faces. As a result of using this model, it was possible to identify and monitor points of interest in thermographic images in an uncontrolled environment.