| Abstrakt: |
Skin cancer is a serious global health issue, and early and accurate categorization of skin lesions is crucial for effective treatment and therapeutic decision-making. However, automated skin cancer classification remains challenging due to issues such as imbalanced and limited training data, model stability, and cross-domain adaptability. Recently, Convolutional Neural Networks (CNNs) have gained popularity for skin cancer classification, showing promising results. However, there are limited reviews that specifically address the cutting-edge challenges in this field. In this research paper, we provide a comprehensive summary of the most recent CNN methods for skin cancer classification. We also highlight freely available datasets for skin cancers and provide an overview of three types of dermatological imaging. The successful applications of conventional CNNs in skin cancer categorization are discussed. Moreover, we identify and discuss key challenges in skin cancer classification, including data imbalance, data restriction, domain adaptability, model resilience, and model efficiency, along with potential remedies. This research contributes to a better understanding of the current state of the field and the practical constraints of skin cancer classification and can serve as a valuable resource for researchers and practitioners in the field of dermatology and medical imaging. [ABSTRACT FROM AUTHOR] |
