Identifikasi Penyakit Vitiligo pada Kulit Menggunakan Metode K-Means Clustering dan Convolutional Neural Network (CNN)

Abstract

Vitiligo is a skin pigmentation disorder characterized by the loss of natural skin color due to damage to melanocytes, and can cause psychosocial impacts for sufferers. Early identification is essential to reduce the risk of complications and social stigma, but the general public often experiences obstacles in accessing medical services, especially in remote areas. This study aims to develop a digital image-based vitiligo identification system using the K-Means Clustering method for skin patch segmentation and Convolutional Neural Network (CNN) for skin image classification. The dataset consists of skin images containing vitiligo and non-vitiligo which are processed through preprocessing, segmentation, and model training stages. The results of the study showed that the use of a model with the K-Means Clustering algorithm and Convolutional Neural Network (CNN) was able to detect vitiligo skin lesions with an accuracy rate of 95%. The results showed that the developed system was able to identify the presence of vitiligo accurately and efficiently.