Skin Lesion Detection using Texture Based Segmentation and Classification by Convolutional Neural Networks (CNN)

Abstract

Skin cancer is one of the dangerous cancers like breast cancer, brain tumour, and lung cancer. The detection of a skin lesion is melanoma or nonmelanoma is a very crucial issue. The earlier detection of melanoma is one of the best solutions for this issue. There is a various technique for detecting the skin lesion. Because of the technology advancement earlier detection of the skin lesion is possible. Malignant melanoma is a very harmful melanoma it is the cancerous cell that will lead to growth and that can be a mole in different colours red, black and brown. Skin lesion segmentation from dermoscopic images is a very challenging task nowadays because of the contrast of those images. there are various techniques for detecting the skin cancer base on the characteristics of the images shape, colour, textures. We proposed a system for skin cancer detection using texture-based segmentation and classification using Convolutional Neural Network. GLCM (Gray Level Co-occurrence Matrix) matrix is exacting the features from an image. And used Neural network tool for checking the accuracy of training network. Nowadays Deep Learning technique is very popular for classification of images. CNN is one of the techniques of Deep Learning. The proposed work will help in classification of skin lesion. Model will helpful for dermatologists for classifying melanomas.

Melanoma Skin Cancer and Nevus Mole Classification using Intensity Value Estimation with Convolutional Neural Network

“…The early-stage detection of melanoma skin cancer can be beneficial in curing it [3,11,12,37]. The early diagnosis survival rate of skin cancer is more than 90 percent [2].…”