Development of a Convolutional Neural Network Architecture for Skin Cancer Type Classification

• Roney Nogueira de Sousa IFCE
• Ana Júlia Lopes de Brito UFC

DOI: https://doi.org/10.5753/ercas.2024.238514

Abstract

In this study, we introduced a convolutional neural network to assist in the diagnosis of skin cancer through images of skin lesions. We used a public database, incorporating data augmentation techniques to address the variability of skin characteristics. After 25 training epochs, we achieved promising results, with an accuracy of 92.91%, AUC of 92.59%, precision of 75.76%, and F1-Score of 71.01%.

References

Chollet, F. et al. (2015). Keras. [link].

Esteva, A., Kuprel, B., et al. (2017). Dermatologist-level classification of skin cancer with deep neural networks. Nature, 542(7639):115–118.

Gavrikov, P. (2020). Visualkeras. [link].

Hunter, J. D. (2007). Matplotlib: A 2d graphics environment. Computing in Science & Engineering, 9(3):90–95.

Jain, S., Pise, N., et al. (2015). Computer aided melanoma skin cancer detection using image processing. Procedia Computer Science, 48:735–740.

Sommer, C. (2008). Skin biopsy as a diagnostic tool. Current opinion in neurology, 21(5):563–568.

Van Rossum, G. and Drake, F. L. (2009). Python 3 Reference Manual. CreateSpace, Scotts Valley, CA.

Vestergaard, M., Macaskill, P., Holt, P., and Menzies, S. (2008). Dermoscopy compared with naked eye examination for the diagnosis of primary melanoma: a meta-analysis of studies performed in a clinical setting. British Journal of Dermatology, 159(3):669–676.

Vieira, E. Q. (2022). Comparação entre diferentes modelos de redes neurais convolucionais para classificação de melanoma. Universidade de Brasília.