Klasifikasi Keparahan Jerawat Berdasarkan Ruang Warna Ycbcr Menggunakan Metode Gray Level Co-Occurrence Matrix dan K-Means

Abstract

Acne is a common skin disorder in Indonesia. More than 80% of people experience skin problems in the form of acne in the age range of 12 to 44 years. Proper identification is needed to determine the severity of acne to get the right treatment. The existence of technological developments regarding image processing can facilitate the process of identifying acne only with acne images. Image processing in determining the severity of acne will be much easier if the skin color is well-represented. The skin color representation that is considered close enough to the way the human eye sees color is by converting the image color space into YCbCr color space. In the process of determining the severity of acne, the Gray Level Co-occurrence Matrix (GLCM) method is used because the method can produce complete extraction features. The results of feature extraction from the Gray Level Co-occurrence Matrix (GLCM) method will be classified using the KMeans method. The combination of the completeness of feature extraction obtained using the GLCM method and the K-Means algorithm makes the classification process run fast, even with a large number of datasets. The results of feature extraction performed using MATLAB produce centroids that are used as a reference for the classification process using K-Means. The extraction process using YCbCr images converted into grayscale produces centroids in clusters for mild acne, moderate acne, and severe acne. Out of 15 tested data, 12 were successfully classified into the right cluster with a percentage of 80%.