Klasifikasi Penyakit Kulit pada Wajah Menggunakan Algoritma Probabilistic Neural Network (PNN)

Abstract

Skin disease is an illness that attacks skin cells and can cause symptoms like swelling, redness, and itchiness. Of all the skin on the human body, facial skin is the most sensitive compared to others. Various skin conditions affect the face, and several have similar colors and textures. One cannot trest skin diseases carelessly. Therefore, an app is needed to help identify facial skin disease more easily and quickly. This study utilized the Probabilistic Neural Network algorithm. There were a total of 500 data, which 350 were the five types of skin disease on the face used for training data and 100 for validation, while remaining 50 were for testing data. Before identifying the disease type, the author performed preprocessing using resizing and grayscaling, segmentation using threshold and then used invariant moment for the feature extraction process to achieve seven moment values and classified them using PNN. With this method, the accuracy result value of the skin disease classification process on the face using Probabilistic Neural Network was 88%.