Using Deep Learning Algorithms for Accurate Diagnosis and Outbreak Prediction of Monkeypox

Monkeypox is a viral illness that has been known to affect the humans. It is commonly misidentified as chickenpox due to the similarity of its rashes with chickenpox, resulting in improper treatment and further spread of the disease. Researchers are using some deep learning algorithms to quickly diagnose monkeypox with accuracy as high as possible. These algorithms are trained on images of monkeypox lesion to distinguish them from other diseases, such as chickenpox and smallpox. This approach has shown encouraging results and high accuracy in monkeypox detection. Nonetheless, with further research and development, deep learning algorithms may prove to be valuable tools in the fight against monkeypox and other infectious diseases. The study involved training a CNN using the MobileNetV2 architecture on the Monkeypox Image Dataset. To optimize loading times, Numpy arrays were used for data preparation instead of loading the images directly. The weights from ImageNet have been used to initialize the MobileNet. The Adam optimizer was utilized to incorporate non-linear behavior into the model, while a loss function and metric were used to score the model's performance. The Model checkpoint and Reduce LR On Plateau techniques were employed to preserve the model weights and decrease the rate of learning by 0.5 if the loss in the verification set did not show any improvement after two successive iterations. A batch size of 32 with 20 epochs have been used to train the model. The model's performance was assessed through a detailed analysis of its predictions, including an assessment of precision, recall, and accuracy. This comprehensive evaluation demonstrated that the model exhibited exceptional performance.