Convolutional Neural Network untuk mengklasifikasi tingkat keparahan jerawat

Abstrak

Classification is one of the methods used in medical science, especially for the early detection or classify the disease types. In skin health, classification can be used to predict the type and severity of acne so that the treatment can be determined. This study aims to develop a classification model for the type and severity of acne using Deep Learning with a Convolutional Neural Network (CNN). The labels used in the training data consist of levels 0, 1, and 2, which represent the severity of acne. The classifier model was developed using secondary data from www.kaggle.com with 500 images for each label. The optimizer used in this study was ADAM by comparing the number of epochs starting from 50, 80, and up to 100. The accuracy results in the training data obtained were 0.6363, 0.8783, and 0.9234, respectively.