Developing an early detection model for skin diseases using a hybrid deep neural network to enhance health independence in coastal communities

The object of study is a solution for early skin disease diagnosis by integrating hybrid deep neural networks – EfficientNetB7 for Classification and YOLOv8 for detection. The system is designed to classify five skin conditions: Melanoma, Basal Cell Carcinoma (BCC), melanoma is a type of skin cancer that originates from melanocytes, the cells that produce skin pigment, Melanocytic Nevi (NV) Melanocytic nevus is a mole or dark spot on the skin formed due to the accumulation of melanocytes, Benign Keratosis-like Lesions (BKL) is a term for a group of skin changes that resemble keratosis but are non-cancerous, and Seborrheic Keratoses and other benign tumors to enhance the health diagnostics. The problem to be solved in this study revolves around improving early and accurate skin disease diagnosis, particularly in resource-limited or underserved areas and the lack of Accessible Diagnostic Tools and Low Efficiency of Current Diagnostic Methods. The study highlights EfficientNetB7's classification accuracy at 94 % and YOLOv8's means average precision (mAP) of 0.812 for detection. This hybrid system processes skin images efficiently, providing classification and detection outcomes with consistent performance in multiple tests. The results demonstrate that the EfficientNetB7 model achieved an accuracy of 94 % on test data, while YOLOv8 delivered a detection performance with a mean average precision (mAP) of 0.812. The web-based system efficiently processed skin images and provided classification and detection outcomes. Furthermore, integrating EfficientNetB7 and YOLOv8 allowed the skin disease detection system to classify five different diseases and assess malignancy risk. The systems are portable and can be used with minimal setup, making them practical for real-world diagnostic use. The Scope and Practical applications are designed for accessibility in resource-limited settings. The website-based skin disease detection tool provides a user-friendly platform accessible to the public and healthcare providers, especially in underserved areas. Each application's high accuracy and ease of use make them viable aids in early diagnosis, potentially improving healthcare access