DETEKSI PENYAKIT KULIT MENULAR BERDASARKAN CITRA LESI KULIT MENGGUNAKAN ALGORITMA VISION TRANSFORMER BERBASIS WEBSITE

Mutiara Putri Rafhsanjani Darmawan, . (2025) DETEKSI PENYAKIT KULIT MENULAR BERDASARKAN CITRA LESI KULIT MENGGUNAKAN ALGORITMA VISION TRANSFORMER BERBASIS WEBSITE. Skripsi thesis, Universitas Pembangunan Nasional Veteran Jakarta.

[img] Text
ABSTRAK.pdf
Download (150kB)
[img] Text
AWAL.pdf
Download (2MB)
[img] Text
BAB 1.pdf
Restricted to Repository UPNVJ Only
Download (169kB)
[img] Text
BAB 2.pdf
Restricted to Repository UPNVJ Only
Download (758kB)
[img] Text
BAB 3.pdf
Restricted to Repository UPNVJ Only
Download (578kB)
[img] Text
BAB 4.pdf
Restricted to Repository UPNVJ Only
Download (3MB)
[img] Text
BAB 5.pdf
Download (153kB)
[img] Text
DAFTAR PUSTAKA.pdf
Download (177kB)
[img] Text
RIWAYAT HIDUP.pdf
Restricted to Repository UPNVJ Only
Download (107kB)
[img] Text
LAMPIRAN.pdf
Restricted to Repository UPNVJ Only
Download (6MB)
[img] Text
ARTIKEL KI.pdf
Restricted to Repository staff only
Download (588kB)
[img] Text
HASIL PLAGIARISME.pdf
Restricted to Repository staff only
Download (13MB)

Abstract

The diagnosis of viral skin diseases such as monkeypox, chickenpox, measles, cowpox, and HFMD typically relies on PCR methods that are time-consuming and require advanced laboratory facilities. Similar early symptoms across these diseases often complicate early diagnosis, while shortages of medical personnel and diagnostic tools worsens the situation, particularly in remote areas. This study aims to develop an automated detection model based on Vision Transformer (ViT) algorithm to classify viral skin diseases from skin lesion images and integrate it into a web-based system. The models were trained using ViT-B/16 and ViT-L/16 architectures with AdamW and RAdam optimizers. Results show that ViT-B/16 with RAdam outperformed ViT-L/16 with AdamW, highlighting the critical role of optimizer selection. The best performance was achieved by ViT-L/16 with RAdam, showing a validation loss of 0.2729, along with 93.83% accuracy, 96.81% precision, 92.45% recall, and 94.06% f1-score on test data. The model was integrated into a website to enable real-time detection for users.

Item Type: Thesis (Skripsi)
Additional Information: [No.Panggil: 2110314007] [Pembimbing 1: Ni Putu Devira Ayu Martini] [Pembimbing 2: Yosy Rahmawati] [Penguji 1: Didit Widiyanto] [Penguji 2: Silvia Anggraen]
Uncontrolled Keywords: Viral Skin Diseases, Vision Transformer, Image Classification, Real Time
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisions: Fakultas Teknik > Program Studi Teknik Elektro (S1)
Depositing User: MUTIARA PUTRI RAFHSANJANI DARMAWAN
Date Deposited: 21 Jul 2025 04:56
Last Modified: 21 Jul 2025 04:56
URI: http://repository.upnvj.ac.id/id/eprint/38129

Actions (login required)

View Item View Item