Skin Cancer Image Detection and Classification by CNN based Ensemble Learning

Melanoma is accounted as a rare skin cancer responsible for huge mortality rate. However, various imaging tests can be used to detect the metastatic spread of disease with primary diagnosis or on clinical suspicion. Focus melanoma detection, irrespective its unusual occurrence, that it often misdiagnosed other malignancies leading medical negligence. Sometimes detected only when metastasis has entered bloodstream lymph nodes. So, effective computational strategies early detection are essential. There four principal types two sub types: Superficial spreading, nodular, lentigo, lentigo maligna, Acral lentiginous, and Subungual melanoma. Amelanotic melanoma, one particular type exists in all kinds tones. Classifications classes focused this research. Misclassification errors, overfitting issues improve accuracy, ensemble classifier models, namely Adaboost, random forest, voted ensemble, CNN, Boosted SVM, GMM, have been classification. The results achieve high classification accuracy. imbalanced found six Transfer learning ensembled transfer approaches implemented reduce issues, performances analyzed. Four ML/DL five models investigation. Implementation 19 classifiers analyzed using standard performance metrics such Accuracy, Precision, recall, Mathew’s correlation coefficient, Jaccard Index, F1 measure, Cohen’s Kappa.