Deep Learning Approach to Identify Skin Cancer Diagnosis Using Decoding Predefined

Abstract

Skin cancer is one of the most rapidly spreading illnesses in the world and because of the limited resources available. Early detection of skin cancer is crucial accurate diagnosis of skin cancer identification for preventive approach in general. Detecting skin cancer at an early stage is challenging for dermatologists, as well in recent years, both supervised and unsupervised learning tasks have made extensive use of deep learning. Utilizing transfer learning on five cutting-edge convolutional neural networks, both plain and hierarchical classifiers were developed to distinguish seven mole types using the HAM10000 dataset of dermatoscopic images. Incorporating data augmentation techniques, the DenseNet201 network emerged as the most effective, boasting high accuracy and F-measure with minimized false negatives. Interestingly, the plain model outperformed the hierarchical one, particularly in binary classification distinguishing nevi from non-nevi. The research also outlines an extension employing a UNET segmentation model to precisely identify and segment affected areas, aiding doctors in assessing the extent of skin disease.