Skin Cancer Detection by Feature Extraction using Neural Networks

Abstract

One of the most serious diseases, human cancer is primarily brought on by various molecular changes and genetic instability. Skin cancer is the most prevalent type of cancer in people. We will research and examine them using various segmentation and feature extraction approaches in order to detect skin cancer at an early stage. Due to the high concentration of melanoma-Here, we provide our skin, in the dermis layer of the skin, our focus is on the identification of malignant melanoma skin cancer. Here, we applied our ABCD rule dermoscopy [1] technology for the identification of malignant melanoma skin cancer. In this system, there are several steps for melanoma skin lesion characterization, including image acquisition, pre-processing, segmentation, and feature definition for skin features. Lesion characterization and classification methods are then determined by the lesion's characterization and features. We employed LBP to extract the texture-based features as well as symmetry detection, border detection, color detection, and diameter detection in the feature extraction process for digital image processing. Here, we suggested using a convolutional neural network to distinguish between stages of healthy skin and skin illness.