Neural Network Based Detection of Melanoma Skin Cancer

Skin cancer is the deadliest form of cancers in humans. It is found in various types such as Melanoma, Basal and Squamous cell Carcinoma among which Melanoma is the most unpredictable. Detection of melanoma skin cancer in the earlier stage is very critical. In this paper, we explain the method for the detection of Melanoma Skin Cancer using Segmentation. The input to the system is the Dermoscopic Image and then apply and then by applying novel image processing techniques. The different stages of detection involves- collection of Dermoscopic Images, filtering the images by using Dull Razor filtering for removing hairs and air bubbles in the image, converting to gray scale, contrast enhancement, noise filtering, segmenting the images using Maximum Entropy Threshold. Feature Extraction technique used is Gray Level Co-occurrence Matrix (GLCM). It is a powerful tool for image feature extraction by mapping the gray level co-occurrence probabilities based on spatial relations of pixels in different angular directions. They are mean, standard deviation, Skewness, Kurtosis, Contrast, Energy and Homogeneity. Mean or expected value provides a measure of distribution. The sensitivity and specificity for diagnosis of melanoma achieved by neural network analysis of Raman spectra were 85 % and 99 %, respectively. We propose that neural network analysis of near-infrared Fourier transform Raman spectra could provide a novel method for rapid, automated skin cancer diagnosis on unstained skin samples.