Skin Cancer Detection using Multi ScaleDeep Learning and Transfer Learning

Skin Cancer is on the rise and Melanoma is the most threatening typeamong the skin cancers. Early detection of skin cancer is vital in order toprevent the cancer to be spread to other parts. In this paper a transfer-learning based system is proposed for Melanoma lesions detection. In theproposed system first, the images are preprocessed for removing the noiseand illumination effect. In the next step a convolutional neural networkis trained based on transfer learning using the weights of ImageNet dataset. In the third step the network is fine-tuned to become more specializedfor detecting the Melanoma versus other types of benign cancers. Theproposed system uses the information from the image in 3 stages. In eachstage the focus will be more concentrate on the center on the image wherethe suspicious part is. The results from these parts are combined andapplied to a fully connected neural network. Results shows the superiorityof the proposed methods compare to other state-of-the arts methods.