Skin Lesion Recognition and Classification Based on Deep Learning

Skin cancer is one of the most common cancers in the world, characterized by rapid progression and high lethality. Early diagnosis of skin lesion types can significantly improve patient recovery rates. So, methods of automatic skin lesion recognition are proposed, struggling to achieve high accuracy. In recent years, deep learning has achieved significant results in various research fields. Existing methods for skin lesion classification primarily utilize convolutional neural networks and residual networks, but their accuracies and efficiencies still need improvement. Therefore, we propose a deep learning model that integrates the VIT and SEN et50 architectures for the automatic classification of seven different types of skin lesions. Experimental results demonstrate that the model achieves accuracy of 0.932.