MonkeyPox

Published: 3 April 2026| Version 1 | DOI: 10.17632/st6kggjr23.1
Contributor:
Samuel Egwu

Description

This dataset contains a total of 3,000 skin lesion images across six diagnostic categories, with 500 images per class. The classes include: Monkeypox, Chickenpox, Normal (healthy skin), Acne Level 1 (Acnel1), Acne Level 2 (Acnel2), and Acne Level 3 (Acnel3). The dataset was curated and extended from existing sources to achieve a balanced class distribution, providing a robust foundation for training and evaluating deep learning models for Mpox detection and skin lesion classification. The dataset was used in the development and evaluation of the Mpox-HSAM model, a hybrid deep learning system for early Mpox detection.

Files

Categories

Health

Licence