Klasifikasi Jenis Jerawat pada Data Citra Jerawat Wajah Menggunakan Convolutional Neural Network

Putri, Chatarina Natassya
Qornain, Wafi Dzul
Bamahri, Fakhirah
Yuliastuti, Gusti Eka
Kurniawan, Muchamad

TIN: TERAPAN INFORMATIKA NUSANTARA
Vol 5 No 2 (2024): July 2024

Acne is a condition caused by pilosebaceous inflammation which affects 85% of skin conditions in adolescents and adults. Acne has an impact on the psychological and social health of sufferers. To treat acne, it is necessary to know the right type of acne so that sufferers can treat the type of acne according to how they are treated. This research was carried out to classify the types of acne in facial acne images using the Convolutional Neural Network (CNN) method. Based on previous research, it shows that the use of CNN is considered effective and appropriate in increasing classification accuracy. This research uses a dataset of acne types from Kaggle with a total of 351 data, divided into 5 classes, namely acne fulminans, acne nodules, fungal acne, papules and pustules which will be tested using 2 different optimizers, namely Adam and RMS- prop. From the results of this test, the highest accuracy was 100% using the Adam optimizer and the RMS-prop optimizer test obtained the highest accuracy value of 80%.