CNNs for Automatic Skin Cancer Classification

1. Electrical and Electronics Engineering, University of Tripoli, Tripoli, Libya.
2. Higher Institute of Medical Sciences and Technologies, Algarabulli, Libya.
3. Biomedical Engineering, University of Tripoli, Tripoli, Libya.
* Corresponding author. Tel: (+218)916094184; Email: a.abougarair@uot.edu.ly.
Manuscript submitted July 11, 2024; revised August 6, 2024; accepted August 22, 2024.

Abstract—Skin cancer is a serious public health issue, and successful treatment depends on an early and precise diagnosis. The capacity of Convolutional Neural Networks (CNNs) to automatically learn and extract significant characteristics from skin lesion photos has made them an effective tool for classifying skin cancer. This paper provides an abstract on the use of CNNs in skin cancer classification and discuss the importance of training CNN models on diverse and comprehensive datasets, the architecture of CNNs, and their capability to capture intricate patterns and features in skin lesions. Moreover, we highlight the potential of CNNs in aiding dermatologists in the early detection and diagnosis of skin cancer. Furthermore, we identify several future directions for research, including the expansion of datasets, integration of clinical information, enhancement of model interpretability, exploration of transfer learning and evaluation of robustness against adversarial attacks. Overall, CNNs have demonstrated considerable promise in advancing skin cancer classification, leading to improved diagnostic accuracy and patient care. .

keywords—Artificial intelligence, CNN, Skin cancer.

Cite: Ahmed J. Abougarair, Hesham A. Enheba, Mosab J. Abugarir, Shada E. Elwefati"CNNs for Automatic Skin Cancer Classification," Journal of Advances in Artificial Intelligence vol. 2, no. 2, pp. 218-234, 2024.

Copyright © 2024 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).