Klasifikasi dan Deteksi Penyakit Kulit Menggunakan Metode EfficientNetB7 dan YOLOv8 Berbasis Website

Abstract

Skin diseases are a common health problem experienced by people in Indonesia, especially in areas with limited access to medical services. Early detection of skin diseases is very important, but it is often difficult to do without the help of medical professionals. Therefore, this research aims to develop an artificial intelligence-based diagnosis system that can help detect skin diseases more accurately. Convolutional Neural Network (CNN) has been proven effective in image classification and is implemented in this research. This research presents a solution for early diagnosis of skin diseases by utilizing CNN based on EfficientNetB7 for classification and YOLOv8 for detection. The system is designed to classify five types of skin diseases: Melanoma, Basal Cell Carcinoma (BCC), Melanocytic Nevi (NV), Benign Keratosis-like Lesions (BKL), and Seborrheic Keratoses and other Benign Tumors, and detect whether the disease is cancer or not. The results showed that the EfficientNetB7 model achieved 94% accuracy on the test data, while YOLOv8 showed detection performance with a mean average precision (mAP) of 0.812. The web-based system developed was able to process skin images and provide classification and detection results efficiently, and proved stable in various performance tests. The combination of EfficientNetB7 and YOLOv8 in the early diagnosis system of skin diseases has led to the development of a new system.