Skin Lesion Prediction and Classification using Deep Learning methods

In the past decade, The Skin Cancer is observed to be a most common disease. Eventually it is not considered as a fatal disease if properly treated. The early diagnosis is more important for the proper cure. As per the reports, it is said over 20% of the population in Canada, US are mostly affected. It is also reported that, with respect to the age, the chance of getting affected by the skin cancer is more. The better recovery is seen once the early diagnosis is done. As per the available medical records, the recovery rate is huge and the fatal rate is less. In this research, investigations on identifying the suitable methodology in deep learning technique is carried out. The research gave the optimum solution of efficient skin cancer prediction with an accuracy of 99.6%. The EfficientNetB0 methodology with the Adam optimizer is observed with computationally improved in performance and accuracy when compared with the state of the art methods.