Classification of Different Stages of Eczema Disease using YOLOv8

Skin diseases pose significant challenges in both diagnosis and treatment due to their diverse manifestations and complexities. In recent years, the advancement of deep learning techniques has revolutionized medical image analysis, offering promising avenues for automated detection and categorization of dermatological conditions. The primary focus of this system is the detection of different stages of the skin disease “Eczema” using YOLO model and will be integrated with mobile application. Our methodology entails training the YOLOv8 model on a dataset encompassing 3000 diverse instances of Eczema at various stages. The model undergoes fine-tuning to discern distinctive visual patterns corresponding to different phases of the skin condition. Subsequently, the trained YOLOv8 model is deployed for real-time inference on new images, enabling prompt and reliable identification of Eczema stages. Post-collection, the dataset undergoes labelling to categorically define the various stages of eczema. Performance evaluation of this model demonstrates a commendable accuracy rate of approximately 81.9%, reflecting its robustness in identifying different stages of Eczema.