Model Klasifikasi Citra Penyakit Monkeypox Berbasis Ekstraksi Fitur GLCM dan Algoritma SVM

Abstract

Monkeypox disease is an infectious disease that requires early detection to support effective and rapid treatment. This study aims to develop a Monkeypox disease image classification model with a texture-based approach using the Gray Level Co-occurrence Matrix (GLCM) method and the Support Vector Machine (SVM) classification algorithm. The dataset used is the Monkeypox Skin Images Dataset (MSID) with a total of 3,200 images, consisting of 1,600 Monkeypox infected images and 1,600 normal skin images. All images go through preprocessing stages such as resizing, converting to grayscale, normalization, and median filtering. Furthermore, GLCM texture feature extraction (contrast, energy, correlation, homogeneity) is carried out and the results are used as input for classification using SVM. The evaluation was carried out by testing four SVM kernels: linear, polynomial, RBF, and sigmoid. The test results showed that the RBF kernel gave the best performance with an accuracy of 80%, followed by the linear kernel (73%), sigmoid (68%), and polynomial (65%). These findings prove that the combination of GLCM texture features with SVM algorithm, especially RBF kernel, has strong potential to support automatic diagnosis of Monkeypox disease based on medical images.