Artificial Intelligence Framework Based on Dconvnet for Skin Cancer Detection.

Abstract

Abstract — From the past and current few years, the furthermost common type of cancer is skin cancer out of all the cancers of human. Every year, more than 1 million new cases are occurring in a predictable situation. Different research methods have been proposed by researchers to detect the skin cancer. To classify normal and abnormal form of skin cases, a system for screening is discussed in this article which is developed with a framework of artificial intelligence with deep learning convolutional neural networks. It is focusing on hybrid clustering for segmentation on skin image and crystal contrast enhancement. Initially filtering and enhancement algorithms will be applied, later segmentation will be done followed by Feature’s extraction and classification are included in the developing process. Each step is designed with effective algorithms to achieve the higher accuracy for the detection of cancer. Images are divided into sub-bands to extract the features and those are the inputs for classification system to find either image is cancerous or noncancerous. The different state of art methods is compared with the method proposed in this article.