Transfer Learning Enabled CAD System for Monkey Pox Classification

Skin disease diagnosis using modern artificial intelligence (AI) approaches proved impressive results in classifying diseases compared to dermatologists. Employing AI techniques in computer-aided diagnosis systems will helps in self-identification using intelligent nodes in healthcare facilities. Convolutional Neural Networks (CNN) have been employed by many previous researchers for skin lesion classification. In this work, the use of several Transfer Learning will be introduced for classifying monkey pox from other skin diseases. Even though deep learning (DL) and Transfer Learning (TL) have advantages over a dermatologist, they also have significant weaknesses, such as the potential to incorrectly identify photos in some cases. The achieved results conclude the best environment and techniques for diagnosing monkey pox from other skin diseases. Accuracy, loss, precision, recall, and F1 score will be used to assess the suggested system’s performance. Using TL in Keras and computer vision models, the system successfully diagnoses 3 distinct types of skin illnesses with an accuracy rate of 99.3 %. The best model was InceptionResNetV2 with exponents with epoch 50 and Adam optimizer.