Supervised Classification of Dermoscopic Images using Optimized Fuzzy Clustering based Multi-Layer Feed-Forward Neural Network

Medical image segmentation is the utmost imperative procedure to assist in the conception of the structure of prominence in medical images. Malignant melanoma is the most recurrent type of skin cancer but it is remediable, if diagnosed at a premature stage. Dermoscopy is a non-invasive, diagnostic tool having inordinate possibility in the prompt diagnosis of malignant melanoma, but their interpretation is time overwhelming. Numerous algorithms were established for classification and segmentation of Dermoscopic images. This Research work proposes the tasks of extracting, classifying and segmenting the Dermoscopic image using a more Efficient supervised learning approach, I.e., Multi-Layer Feed-forward Neural Network for more accurate and computationally efficient segmentation. The features are extracted from the Dermoscopic image using Genetically Optimized Fuzzy C-means clustering approach and these accurate features are used to train the multilayer classifier. The trained network are used for segmentation of malignant melanoma from the skin. The results will be compare with the ground truth images and their performance is evaluate after completion of work. The results will be in form of various validation parameters and should outperform the existing supervised learning approaches.