View in EDS

Novel vision transformer and data augmentation technique for efficient detection of monkeypox disease.

Saved in:
Bibliographic Details
Title: Novel vision transformer and data augmentation technique for efficient detection of monkeypox disease.
Authors: Alarfaj, Aisha Ahmed, Ahmad, Salman, Hakeem, Abeer M., Alabdulqader, Ebtisam Abdullah, PERO, Chiara, Alsubai, Shtwai, Innab, Nisreen, Ashraf, Imran
Source: Multimedia Tools & Applications; Aug2025, Vol. 84 Issue 27, p31955-31973, 19p
Subject Terms: MONKEYPOX, DATA augmentation, DEEP learning, EARLY diagnosis, MACHINE learning, CUTANEOUS manifestations of general diseases, TRANSFORMER models, COMPUTER-aided diagnosis
Abstract: Recent technological advancements have paved the way for the optimization of medical processes, particularly automated disease detection. Moreover, the adoption of machine learning (ML) has greatly helped in automating disease detection. Such approaches can detect various diseases early, enabling timely treatment to save countless lives. Early and accurate diagnosis is very important for diseases like monkeypox, to curb its spread. Monkeypox is a viral disease caused by double-stranded DNA and can be transmitted through close contact with infected humans or animals. It's early identification and accurate lesion diagnosis are critical to contain the disease. This study proposes an automated approach to optimize the diagnosis of monkeypox disease using a novel vision transformer, which is utilized due to its effectiveness for feature extraction. The Proposed approach's efficiency and accuracy are tested on a public benchmark dataset comprising a variety of skin lesions of different ages and genders. In addition, data augmentation involves rotation, scaling, and flipping thereby enhancing the density of the training data set for better generalization of ML models. Experiments involve binary, as well as, multi-class classification. For the binary class, the proposed model achieves an accuracy of 97.63%, outperforming traditional ML and deep learning (DL) techniques. In the case of multi-class classification with monkeypox, measles, normal, HFMD, cowpox, and chickenpox classes, the proposed model archives an accuracy of 90.61% while precision, recall, and F1 scores are 91.39%, 89.17%, and 90.28%, respectively. Furthermore, the proposed approach shows average accuracy, precision, recall, and F1 scores of 97.54%, 96.19%, 95.16%, and 95.67%, respectively for five-fold cross-validation. Experiments demonstrate that the combination of data augmentation techniques and the vision transformer model significantly optimizes diagnostic performance. In brief, advanced DL architectures with image data-augmentation strategies can help achieve optimal processes for diagnosing diseases like monkeypox, and avoid widespread outbreaks. [ABSTRACT FROM AUTHOR]
Copyright of Multimedia Tools & Applications is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Complementary Index
Description
Abstract:Recent technological advancements have paved the way for the optimization of medical processes, particularly automated disease detection. Moreover, the adoption of machine learning (ML) has greatly helped in automating disease detection. Such approaches can detect various diseases early, enabling timely treatment to save countless lives. Early and accurate diagnosis is very important for diseases like monkeypox, to curb its spread. Monkeypox is a viral disease caused by double-stranded DNA and can be transmitted through close contact with infected humans or animals. It's early identification and accurate lesion diagnosis are critical to contain the disease. This study proposes an automated approach to optimize the diagnosis of monkeypox disease using a novel vision transformer, which is utilized due to its effectiveness for feature extraction. The Proposed approach's efficiency and accuracy are tested on a public benchmark dataset comprising a variety of skin lesions of different ages and genders. In addition, data augmentation involves rotation, scaling, and flipping thereby enhancing the density of the training data set for better generalization of ML models. Experiments involve binary, as well as, multi-class classification. For the binary class, the proposed model achieves an accuracy of 97.63%, outperforming traditional ML and deep learning (DL) techniques. In the case of multi-class classification with monkeypox, measles, normal, HFMD, cowpox, and chickenpox classes, the proposed model archives an accuracy of 90.61% while precision, recall, and F1 scores are 91.39%, 89.17%, and 90.28%, respectively. Furthermore, the proposed approach shows average accuracy, precision, recall, and F1 scores of 97.54%, 96.19%, 95.16%, and 95.67%, respectively for five-fold cross-validation. Experiments demonstrate that the combination of data augmentation techniques and the vision transformer model significantly optimizes diagnostic performance. In brief, advanced DL architectures with image data-augmentation strategies can help achieve optimal processes for diagnosing diseases like monkeypox, and avoid widespread outbreaks. [ABSTRACT FROM AUTHOR]
ISSN:13807501
DOI:10.1007/s11042-024-20456-9