Skin Cancer Classification of Dermoscopic Images using Deep Learning

Pigmented skin lesions must be identified in order to detect hazardous skin-related disorders, especially cancer. To improve the correctness of skin cancer recognition, image detection strategies and classification abilities can be utilized. Based on the HAM10000 dataset, which contains 10015 photos, this research paper makes use of the dataset. An augmentation has been carried out on a subset of the dataset in the proposed work. Compared to an approach with &without data augmentation, is more likely to learn unique traits and attributes. To make the model more accurate, data augmentation is often used. However, until the model is strong, it will not be able to produce meaningful results when tested with the testing data. This work's model is robust since it makes use of the k-fold cross-validation technique. Machine learning methods and models from Convolutional Neural Networks have been evaluated for their classification accuracy. After comparing the suggested work's use of several machine learning methods, we found that Convolutional Neural Network produced the best results in terms of accuracy. We achieved a maximum accuracy of 95.53% using the CNN in the suggested setup. Early detection of seven types of skin illness is made easier by the suggested work, which may be authenticated and treated properly by medicinal professionals.