Klasifikasi Penyakit dan Kelainan Bentuk Kuku Manusia Menggunakan Convolutional Neural Network

Abstract

Nails are one of the body parts that have an important role, because nails can provide signals of a disease starting from the color, shape, and size of the nails. This study aims to identify diseases and abnormalities of human nail shape using digital images. This study uses seven classes of diseases, namely beausline, clubbing, koilonychia, yellow nail, white nail, onychomycosis, and normal nail. To identify diseases and abnormalities of human nail shape, the method used is Convolutional Neural Network with VGG-16 architecture. The results obtained from this study are an accuracy of 92.9% from a total dataset that has been augmented as much as 3900 data, for an average precision measurement result of 86.1%, an average recall result of 85.8%, an average f-1 score result of 85.8%.