Deep Learning Approaches for Precision Detection, Optimized Classification, and Diagnostic Accuracy Enhancement in Pox Disease Using Transfer Learning

Although COVID-19 had a critical need for rapid and accurate disease detection, while zoonotic infections and pox disease were significant health risks it also took slower transmission than COVID-19. The traditional pox detection is done through manual inspection which is inaccurate, so this increased the need for efficient and accurate diagnostic methods. Polymerase Chain Reaction is a laboratory test used to recognize the small amount of pox virus that is detected but costly and time-consuming. The recent development in Artificial intelligence and deep learning has revolutionary medical diagnostics by high-accuracy classification. This research proposes a deep learning-based classification model using the Xception convolutional neural network (CNN) to classify monkeypox, cowpox, chickenpox, and normal skin conditions which uses transfer learning, and pre-trained weights to enhance feature extraction and classification accuracy. Experimental analysis results that indicate the proposed model achieves a classification accuracy of 95.67%. Additional performance metrics, including precision 95.85%, F1-score 95.81%, Sensitivity 95.77%, and specificity 98.58% validate its effectiveness. This contributes to the advancements in AI-driven healthcare solutions.