Monkeypox and Chickenpox Identification Using Machine Learning Approach

View at publisher Summarise with VIRAT
Authors: Nur Faizatul Azira Mohd Azhar, S.N.M. Halim, Sofianita Mutalib, R. Badlishah Ahmad
Cited by: 0
References: 19

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The world is still recovering from the trauma and overwhelming situations of the COVID-19 outbreak. Yet, the recent outbreaks of monkeypox in 2022 have raised serious concerns among public health authorities. Skin lesions can be a factor to determine diseases because they show a visual appearance that could be a sign of any disease. However, there are some diseases that have developed similar skin lesion appearances. Monkeypox and other viral infections, such as chickenpox, exhibit similar dermatological and physical symptoms, which can result in diagnostic confusion or misidentification. Such misdiagnoses may delay appropriate treatment, hinder outbreak control measures, and increase the risk of transmission, highlighting the urgent need for accurate differentiation. This study aimed to build a prototype to identify monkeypox and chickenpox disease based on skin lesions using a machine learning approach. This study involved techniques of constructing and evaluating Convolutional Neural Networks (CNNs) models such as MobileNetV2, Custom CNN and NASNet. The dataset was collected from Kaggle.com website and consists of patients' skin lesion images. These models underwent extensive experiments with various hyperparameter tuning, such as different split validations, optimizers, number of epochs and batch sizes. The model with the highest testing accuracy was selected and developed into a prototype. The NASNet model, with its highest testing accuracy of 99.88%, outperformed other models in experimentation. A prototype integrated with this model was tested to measure its functionality and performance, resulting in 90% accuracy. This prototype holds potential to aid the public in identifying such diseases effectively. Future work may involve expanding the prototype scope to recognize a broader range of skin conditions by enhancing accuracy through larger and more diverse datasets while considering integrating explainable artificial intelligence (XAI) into the prototype.

Fields of Study & Subjects

Artificial Intelligence in Healthcare Advanced Neural Network Applications Poxvirus research and outbreaks

Publication details

Journal / source
International Conference on Cognitive Systems and Computer Interaction (ICoSCI)
Pages
1–6
Publication date
15 January 2026
Language
EN