Monkeypox Disease Detection using Deep Learning Techniques

Objectives: To strengthen diagnostic precision and slow the transmission of the illness by creating an AI-based system to speed up monkeypox detection via an improved VGG-19 Convolutional Neural Network (CNN) system. Method: In order to maximize performance, bottleneck layers were integrated into a CNN model centered on the VGG-19 design. An openly accessible dataset of monkeypox photos was used to train and evaluate the algorithm. Accuracy, F1 Score, Precision, and Recall criteria were used to assess performance. Findings: With 95% accuracy, the VGG-19-based model showed promise in monkeypox detection. The model has 94.8% accuracy, 95.1% recall, and 94.9% F1 score, showing a balanced result over all major criteria. These findings beat the original CNN model, and this had 90% accuracy, and various VGG-19 variations with dropout layers and other learning rates. The suggested model outperformed state-of-the-art models such as the updated VGG-16 by Ahsan et al. (2022) & MonkeyNet by Bala et al. (2023), demonstrating its strength and dependability for medical diagnostics usage. Novelty: A customized VGG-19 CNN model using bottleneck layers created especially to enhance classification of images efficiency for monkeypox detection is presented in the paper, providing a useful tool towards rapid diagnosis and epidemic containment. Keywords: Monkeypox Detection, Deep Learning, CNN, VGG-19, Medical Image Processing, Transfer Learning, Image Classification