Machine learning approach for monkeypox detection system from medical images

Nazaimi, Nurin Khairina and Mansor, Hasmah and Gunawan, Teddy Surya and Md. Yusoff, Nelidya (2025) Machine learning approach for monkeypox detection system from medical images. In: IEEE International Conference on Smart Instrumentation, Measurement and Applications (ICSIMA), 10-11 September 2025, Royale Chulan, Kuala Lumpur Malaysia.

PDF - Published Version Restricted to Registered users only Download (1MB) | Request a copy

Abstract

The rapid global spread of monkeypox, including outbreaks in non-endemic regions, has raised public health concerns and highlighted the need for rapid, accessible and reliable diagnostic tools. This need is especially critical in resource-limited settings, where conventional methods such as Polymerase Chain Reaction (PCR) face limitations due to high cost, equipment dependency and time consumption. This study proposes a deep learning-based multiclass classification system using GoogLeNet to detect monkeypox from medical skin images. The Monkeypox Skin Image Dataset (MSID), consisting of 770 images across four categories: monkeypox, chickenpox, measles, and normal, is used for model training and evaluation. Through transfer learning and image preprocessing technique, the proposed model achieved an overall accuracy of 91.56%, with precision, recall, and F1-score at 91.64%, 91.56% and 91.43% respectively. Comparative analysis with EfficientNet- B0 and ResNet-18 demonstrates that GoogLeNet outperforms both in terms of generalization and class-wise detection accuracy, which suggest its suitability as a lightweight and effective tool for early monkeypox diagnosis. Furthermore, GoogLeNet was also evaluated against models reported in a benchmark study, including VGG16, ResNet50, MobileNetV1, InceptionV3, Xception, and the custom MonkeyNet architecture. GoogLeNet achieved competitive results with minimal fine-tuning, highlighting its practicality and strong performance despite being a standard open-source model.

Item Type:	Proceeding Paper (Plenary Papers)
Uncontrolled Keywords:	Monkeypox, Deep Learning, Medical Image Classification, GoogLeNet, Convolutional Neural Networks (CNNs)
Subjects:	T Technology > T Technology (General)
Kulliyyahs/Centres/Divisions/Institutes (Can select more than one option. Press CONTROL button):	Kulliyyah of Engineering Kulliyyah of Engineering > Department of Electrical and Computer Engineering
Depositing User:	Dr Hasmah Mansor
Date Deposited:	20 Nov 2025 16:22
Last Modified:	20 Nov 2025 16:25
Queue Number:	2025-11-Q062
URI:	http://irep.iium.edu.my/id/eprint/124481

Actions (login required)

View Item