Melanoma Skin Cancer Detection using Deep Learning

Among the deadliest diseases is skin cancer, particularly melanoma. There is a lot of overlap between various skin lesions like melanoma and moles in the color skin photos. The challenge of discovery and diagnosis is made more challenging by nevus. To prevent unnecessary effort, time, and human life loss, an accurate automated method for skin lesion classification is required. This study offers a technique for categorizing skin lesions automatically. our system relies on the predictions of algorithm called CNN (Convolutional Neural Network). CNN are mainly used for finding patterns in images to identify classes, and categories and it learns from the data. The source of dataset is International Skin Imaging Collaboration (ISIC) and we chose to work with total 2836 images, from this experiment we got the accuracy of 85 to 91%