Skin Cancer – Melanoma Detection in Skin Images Using Local Binary Pattern (LBP) and GLCM

The skin properties like skin dryness, fungus and allergic symptoms i.e. etching kind of problem that may led to starting symptoms of malignant melanoma skin cancer. The correct identification of skin spots based on certain features is the key steps in detecting the skin cancer disease in advance. To improve the accuracy level, a k-means clustering is proposed followed by local binary pattern. This not only clearly detects the melanoma but also segment the cancerous part from the back ground. Further, the image is confirmed by using the local binary pattern in order to do the dimensional analysis of the skin cancer. The algorithm is tested on different skin image data base covering different stages of skin cancer and b=normal images. The results very accurate and later stage could be predicted in consultation with medical practioner. The prime concern in the presented work is on extracting the skin image features in textural domain as well as radial domain i.e. area, perimeter and standard deviation of radii. This enables in analyzing the cancer spot analysis and guides for the direction of spread of the cancer. This is the vital information where the skin expert may get vital information at fine accuracy.