Skin Cancer Detection

Skin Cancer Detection

Author Name : Mr. K. Sanjeeviah, K. Rajesh Kumar, A. Bharath Chandra, D. Sai Chand, G. Mahindra Yadav

ABSTRACT

Unlike benign illnesses, skin cancer is a prevalent form of human malignancy that may be visually diagnosed beginning with clinical screening and culminating with histological testing. The close closeness of the picture features makes it difficult to automatically identify skin lesion images effectively. We employed two techniques for picture recognition in this work. The first method makes use of convolution neural networks and has promise as a possible classifier for skin lesions. (CNN). This study uses Custom CNN, a deep learning convolution neural network (DLCNN) that converts pictures to class labels, to deliver an expert-level classification of skin cancer. It is demonstrated that a classifier with a single CNN can automatically distinguish between photos of healthy and cancerous skin. For the second strategy, we used the Support Vector Machine. Its supervised learning approach divides data into several groups based on a dividing hyperplane. The only network inputs utilised are disease names and image pixels. The model's performance has been evaluated using images from the two labels that were not trained. This model recognises the most common skin cancers and can be upgraded with an endless number of new pictures. The DLCNN was trained using the Custom model, and it displayed accurate classification of the benign and malignant skin categories.