Automated Skin Lesion Classification Using Ensemble of Deep Neural Networks in ISIC 2018: Skin Lesion Analysis Towards Melanoma Detection Challenge

In this paper, we studied extensively on different deep learning based to detect melanoma and skin lesion cancers. Melanoma, a form of skin cancer is very threatening to health. Proper diagnosis of at an earlier stage is crucial for the success rate of complete cure. images with Benign and malignant forms of skin cancer can be by computer vision system to streamline the process of skin cancer. In this study, we experimented with various neural networks which recent deep learning based models like PNASNet-5-Large, 2, SENet154, InceptionV4. Dermoscopic images are properly and augmented before feeding them into the network. We tested our on International Skin Imaging Collaboration (ISIC) 2018 challenge. Our system has achieved best validation score of 0. 76 for-5-Large model. Further improvement and optimization of the proposed with a bigger training dataset and carefully chosen hyper-parameter improve the performances. The code available for download at: //github. com/miltonbd/ISIC₂018classification