Manuscript received January 4, 2024; revised March 4, 2024; accepted March 11, 2024; published July 25, 2024

Abstract—The purpose of this research is to investigate the utilization of hybrid models in dermatological diagnostics and to demonstrate the potential of these models to advance medical picture classification capabilities. The study presents BiT-EfficientNet, a novel hybrid model developed specifically for the precise classification of monkeypox lesions in skin images. By combining EfficientNet B6 and Big Transfer (BiTM- R50x1), the model demonstrates exceptional performance in recognizing patterns and managing visual features. BiTEfficientNet demonstrates superior performance compared to existing models, achieving a precision of 98.25%, recall of 95.48%, F1-Score of 96.84%, and accuracy of 96.86%. It is positioned as a strong contender through comparative analysis. A highly accurate model is achieved through careful parameter optimization, resulting in a training accuracy of 99.14%. Assessing resilience through empirical means validates it. The findings have a significant impact on increasing diagnostic accuracy for illnesses like monkeypox, which can result in prompt interventions in professional medical professionals’ healthcare.

Keywords—monkeypox disease, deep learning, ensemble learning, image processing, skin lesion detection

Cite: Sharia Arfin Tanim, Kazi Tanvir, Al Rafi Arnob, Md. Hasibur Rahman, Tasmia Binte Munir Maisha, and Kamruddin Nur, "Enhancing Monkeypox Diagnostics: Exploring the Potential of EfficientNet and Big Transfer," Journal of Image and Graphics, Vol. 12, No. 3, pp. 250-258, 2024.

Copyright © 2024 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.