Skin Cancer Classification Using CNN

Skin cancer must need to be detected so that further spread of skin cancers, such as melanoma and focal cell carcinoma can be prevented. There are a lot of things which are having bad impacts on the accuracy of detection. The usage of image processing as well as machine vision in the realm of healthcare and medical applications has increased significantly in the recent scenario. In this paper, we are using convolutional neural networks (CNN) to categorize different forms of skin cancer. Seven different kinds of skin cancer are included in the Skin Cancer Detection dataset which is used to create the proposed system. The CNN design is made up of numerous convolutional layers with max-pooling and batch normalization, which come after a layer which are fully connected, and a SoftMax which is an output layer for classification. Our model has performed well and outperformed conventional machine learning classifiers with an accuracy of more than 95% in the case of almost all the optimizers. Building a CNN model to accurately diagnose skin cancer with a precision of above 80%, lowering its false negativity rate (FNR) in prediction to below 10%, and performing data visualization are some of our goals for this project. The paper shows how CNNs have the potential to effectively classify different types of skin cancer, which can help with the identification and enhancement in a patient’s outcome.