A Deep Neural Network Model for Effective Diagnosis of Melanoma Disorder

Skin cancer is one of the most critical diseases, and melanoma is the most widely observed skin cancer occurring in major population. Hence, early detection and treatment of melanoma disease are very important. This research study presents the classification of melanoma images using deep neural network. Principal component analysis (PCA) was used as a feature extraction algorithm in this study. Wavelet transform is also applied to remove noise and inconsistencies in the dataset. Performance was compared with various other classifiers like SVM, RBS, random forest and Naive Bayes. Deep neural network gave an optimal performance with performance indicators. Classification accuracy, precision, recall and F-score value recorded with deep neural network are 98.4%, 97.8%, 98.5% and 98.1%, respectively. Hence, our proposed work can be an effective classification framework in categorization and diagnosis of melanoma disease.