Automated Segmentation of Lesions Associated with Tinea Pedis Disease

Tinea pedis, commonly referred to as athlete’s foot, remains a prevalent dermatophytic infection affecting the feet. The use of AI-driven healthcare solutions has attracted a lot of attention lately, especially when it comes to the diagnosis and treatment of diseases. This research aims to enhance the current understanding and methods by introducing an automated segmentation technique designed to identify and outline lesions caused by Tinea pedis. Central to this study is the implementation of advanced segmentation models, including UNet and YOLOv9, to achieve accurate and efficient lesion segmentation. The research methodology encompasses the creation of comprehensive datasets through a combination of publicly available images and patient samples from skin clinic. Augmentation techniques and meticulous manual annotation procedures are employed to enhance dataset quality and facilitate model training. Our dataset will be made publicly accessible.Important stages of the study include preprocessing the data, building the model, training, evaluating, and optimizing it. Noteworthy emphasis is placed on the customization of annotation strategies tailored to the requirements of each segmentation model. Through iterative experimentation and parameter tuning, the models are refined to ensure optimal performance in lesion identification and segmentation. The outcomes of this research showcase the potential of AI-driven segmentation models in augmenting diagnostic processes for Tinea pedis. By automating lesion detection and delineation, these models offer a promising avenue for enhancing the efficiency and accuracy of disease diagnosis. The U-Net model demonstrated an IOU of 94.28% and a precision of 95.09%, whereas the YOLOv9 model achieved an IOU of 85.15% and a precision of 89.38%. Moreover, the scalability of the proposed approach lays the groundwork for its application in the broader domain of dermatological diagnostics, thereby fostering advancements in patient care and clinical decision-making.