IMAGE PROCESSING TECHNIQUES FOR MELANOMA SKIN CANCER DETECTION AND CLASSIFICATION

Abstract

Skin cancer is one of the most dangerous types of cancer that are diagnosed in people. The most predominant and possibly hazardous kind of cancer in people is skin cancer. Especially, melanoma skin cancer has a high casualty rate. Powerful treatment relies basically upon early detection. Melanoma is regularly recognized through agonizing, tedious biopsies. An early melanoma conclusion sister computer-upheld detection approach is displayed in this assessment. In this audit, we explicitly feature melanoma dangerous cells utilizing image data and recommend two strategies for identifying skin cancer. Convolutional neural networks, such as the AlexNet, LeNet, and VGG-16 models, are used in the main approach. We integrate the model with the highest degree of accuracy into flexible and online applications. The second method classifies images as normal, dangerous, or safe after image processing by using feature limitations and backing vector machines with a default RBF portion. The CNN model is offered as an accessible internet application, made possible with the aid of Django and Android Studio.