A Hybrid Approach Combining CNN and Ensemble Algorithms for Dermoscopic Image Classification

• Pedro A. A. Soares Centro Universitário FAG
• Leandro A. Ensina UTFPR
• Juliano H. Foleis UTFPR

DOI: https://doi.org/10.5753/eniac.2025.12460

Resumo

Este trabalho apresenta uma abordagem híbrida para a classificação de lesões cutâneas pigmentadas, incluindo tipos de câncer de pele, em imagens dermatoscópicas. A técnica combina redes neurais convolucionais (CNNs) como extratoras de características e algoritmos ensemble como classificadores, além da introdução de duas etapas distintas de pré-processamento para aumento de dados. A primeira etapa ocorre antes do treinamento das CNNs, enquanto a segunda é aplicada antes do treinamento dos classificadores. Os experimentos realizados com a base HAM10000 evidenciam a eficácia do método, alcançando F1-Scores gerais superiores a 80%. Além disso, o estudo aponta direções para trabalhos futuros, visando o aprimoramento do método.

Referências

Afza, F., Sharif, M., Khan, M., Tariq, U., Yong, H.-S., and Cha, J. (2022). Multiclass skin lesion classification using hybrid deep features selection and extreme learning machine. Sensors, 22(3):799.

Ajiboye, A. O. (2024). Hybrid skin lesion detection integrating cnn and xgboost for accurate diagnosis. International Journal of Computer (IJC), 53(1):14–71.

Alenezi, F., Armghan, A., and Polat, K. (2023). A multi-stage melanoma recognition framework with deep residual neural network and hyperparameter optimization-based decision support in dermoscopy images. Expert Systems with Applications, 215:119352.

Benjdira, B., Ali, M., Koubaa, A., Ammar, A., and Boulila, W. (2024). Dm-ahr: A self-supervised conditional diffusion model for ai-generated hairless imaging for enhanced skin diagnosis applications. Cancers (Basel), 16(17):2947. Published: 2024 Aug 23.

Benyahia, S., Meftah, B., and Lézoray, O. (2022). Multi-features extraction based on deep learning for skin lesion classification. Tissue and Cell, 74:101701.

Brancaccio, G., Balato, A., Malvehy, J., Puig, S., Argenziano, G., and Kittler, H. (2024). Artificial intelligence in skin cancer diagnosis: A reality check. Journal of Investigative Dermatology, 144(3):492–499.

Buslaev, A., Iglovikov, V. I., Khvedchenya, E., Parinov, A., Druzhinin, M., and Kalinin, A. A. (2020). Albumentations: Fast and flexible image augmentations. Information, 11(2).

Chang, C.-C., Li, Y.-Z., Wu, H.-C., and Tseng, M.-H. (2022). Melanoma detection using xgb classifier combined with feature extraction and k-means smote techniques. Diagnostics, 12(7):1747. Published: 19 July 2022.

Chollet, F. (2017). Xception: Deep learning with depthwise separable convolutions. In 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 1800–1807.

Cruz, R. M., Sabourin, R., and Cavalcanti, G. D. C. (2018). Dynamic classifier selection: Recent advances and perspectives. Information Fusion, 41:195–216.

Khan, M. A., Sharif, M., Akram, T., Damaševičius, R., and Maskeliūnas, R. (2021). Skin lesion segmentation and multiclass classification using deep learning features and improved moth flame optimization. Diagnostics, 11(5):811. Published: 29 April 2021.

Menegola, A., Fornaciali, M., Pires, R., Bittencourt, F. V., Avila, S., and Valle, E. (2017). Knowledge transfer for melanoma screening with deep learning. In 2017 IEEE 14th International Symposium on Biomedical Imaging (ISBI 2017), pages 297–300.

Roy, A., Cruz, R. M., Sabourin, R., and Cavalcanti, G. D. (2018). A study on combining dynamic selection and data preprocessing for imbalance learning. Neurocomputing, 286:179–192.

Shetty, B., Fernandes, R., Rodrigues, A. P., Chengoden, R., Bhattacharya, S., and Lakshmanna, K. (2022). Skin lesion classification of dermoscopic images using machine learning and convolutional neural network. Scientific Reports, 12(1):18134.

Spolaôr, N., Lee, H. D., Takaki, W. S. R., Coy, C. S. R., and Wu, F. C. (2024). Avaliação de variações da rede profunda efficientnet em bases dermoscópicas. Journal of Health Informatics, 16(especial).

Sá, J., Ensina, L., and Jeronymo, D. (2024). Aplicação de redes de aprendizado profundo e algoritmos de aprendizado de máquina para classificar imagens de câncer de pele. In Anais do XXIV Simpósio Brasileiro de Computação Aplicada à Saúde, pages 651–656, Porto Alegre, RS, Brasil. SBC.

Tschandl, P., Rosendahl, C., and Kittler, H. (2018). The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions. Scientific Data, 5(1):180161.

Yanchatuña, O. P., Pereira, J. P., Pila, K. O., Vásquez, P. A., Veintimilla, K. S., Villalba-Meneses, G. F., Alvarado-Cando, O., and Almeida-Galárraga, D. (2021). Skin lesion detection and classification using convolutional neural network for deep feature extraction and support vector machine. International Journal on Advanced Science, Engineering and Information Technology, 11(3):1260–1267.

Zanddizari, H., Nguyen, N., Zeinali, B., and Chang, J. M. (2021). A new preprocessing approach to improve the performance of cnn-based skin lesion classification. Medical & Biological Engineering & Computing, 59(5):1123–1131. Epub 2021 Apr 26.