Atlas-Guided U-Net++ with EfficientNetB5 for Automatic Pancreas Segmentation in Abdominal CT Scans

• Felipe R. S. Teles UFMA
• Neilson P. Ribeiro UFMA / IFMA
• Luana B. da Cruz UFCA
• Geraldo B. Júnior UFMA
• Anselmo C. de Paiva UFMA
• João O. B. Diniz UFMA / IFMA
• Omar A. C. Cortes IFMA

DOI: https://doi.org/10.5753/sbcas.2025.6940

Resumo

Pancreas segmentation in abdominal computed tomography images is challenging due to the organ’s variability in shape, size, and position. This work proposes an automatic segmentation method based on a 2D Convolutional Neural Network (CNN) approach, consisting of three steps: (1) filtering non-pancreas slices using a CNN, (2) region of interest detection via a probabilistic atlas, and (3) final segmentation with U-Net++ with an EfficientNetB5 backbone. The method achieves a mean Dice coefficient of 78.59% and Recall of 79.12%, with a lower computational cost compared to 2.5 and 3D approaches. Thus, our results stand out among state-of-the-art methods, providing a computationally efficient and accurate solution for diagnosis and treatment planning.

Referências

Antonelli, M., Reinke, A., Bakas, S., Farahani, K., Kopp-Schneider, A., Landman, B. A., Litjens, G., Menze, B., Ronneberger, O., Summers, R. M., et al. (2022). The medical segmentation decathlon. Nature communications, 13(1):4128.

Bray, F., Laversanne, M., Sung, H., Ferlay, J., Siegel, R. L., Soerjomataram, I., and Jemal, A. (2024). Global cancer statistics 2022: Globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: a cancer journal for clinicians, 74(3):229–263.

Daniel, N., Farinella, R., Chatziioannou, A. C., Jenab, M., Mayén, A.-L., Rizzato, C., Belluomini, F., Canzian, F., Tavanti, A., Keski-Rahkonen, P., et al. (2024). Genetically predicted gut bacteria, circulating bacteria-associated metabolites and pancreatic ductal adenocarcinoma: a mendelian randomisation study. Scientific Reports, 14(1):25144.

Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., and Fei-Fei, L. (2009). Imagenet: A large-scale hierarchical image database. In 2009 IEEE Conference on Computer Vision and Pattern Recognition, pages 248–255.

Diniz, J. O., Ferreira, J. L., da Silva, G. L., Quintanilha, D. B., Silva, A. C., and Paiva, A. (2021). Segmentação de coração em tomografias computadorizadas utilizando atlas probabilístico e redes neurais convolucionais. In Anais do XXI Simpósio Brasileiro de Computação Aplicada à Saúde, pages 83–94. SBC.

Fernandes, A. G. S., Braz Junior, G., Diniz, J. O. B., Silva, A. C., and Matos, C. E. F. (2023). Efficientdeeplab for automated trachea segmentation on medical images. In Brazilian Conference on Intelligent Systems, pages 154–166. Springer.

Ferrara, N., Andria, G., Scarpetta, M., Lanzolla, A. M. L., Attivissimo, F., Di Nisio, A., and Ramos, D. (2024). 2d and 2.5 d pancreas and tumor segmentation in heterogeneous ct images of pdac patients. In 2024 IEEE International Symposium on Medical Measurements and Applications (MeMeA), pages 1–5. IEEE.

He, K., Zhang, X., Ren, S., and Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778.

Junior, D. A. D., da Cruz, L. B., and Diniz, J. O. (2024). Classificação da camada lipídica do filme lacrimal usando k-means e deep learning. In Simpósio Brasileiro de Computação Aplicada à Saúde (SBCAS), pages 1–12. SBC.

Kurnaz, E., Ceylan, R., Bozkurt, M. A., Cebeci, H., and Koplay, M. (2024). A novel deep learning model for pancreas segmentation: Pascal u-net. Inteligencia Artificial, 27(74):22–36.

Lee, H., Kim, Y. S., Kim, M., and Lee, Y. (2021). Low-cost network scheduling of 3d-cnn processing for embedded action recognition. IEEE Access, 9:83901–83912.

Li, J., Lin, X., Che, H., Li, H., and Qian, X. (2021). Pancreas segmentation with probabilistic map guided bi-directional recurrent unet. Physics in Medicine & Biology, 66(11):115010.

Long, J., Shelhamer, E., and Darrell, T. (2015). Fully convolutional networks for semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3431–3440.

Neto, C. M. S., Silva, A. L., Pessoa, A. C., Quintanilha, D. B., de Almeida, J. D., Junior, G. B., and Diniz, J. O. (2024). Diagnóstico de tuberculose em imagens de radiografia utilizando cvt. In Simpósio Brasileiro de Computação Aplicada à Saúde (SBCAS), pages 342–353. SBC.

Ronneberger, O., Fischer, P., and Brox, T. (2015). U-net: Convolutional networks for biomedical image segmentation. In Navab, N., Hornegger, J., Wells, W. M., and Frangi, A. F., editors, Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015, pages 234–241, Cham. Springer International Publishing.

Silva, G., Oliveira, F., Diniz, J., Diniz, P., Quintanilha, D., Silva, A., Paiva, A., and Cavalcanti, E. (2021). An automatic method for prostate segmentation on 3d mri scans using local phylogenetic indexes and xgboost. In Anais do XXI Simpósio Brasileiro de Computação Aplicada à Saúde, pages 165–176, Porto Alegre, RS, Brasil. SBC.

Tan, M. and Le, Q. (2019). Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning, pages 6105–6114. PMLR.

Tian, M., He, J., Yu, X., Cai, C., and Gao, Y. (2021). Mcmc guided cnn training and segmentation for pancreas extraction. IEEE Access, 9:90539–90554.

Weston, A. (2020). How to segment a pancreas ct. Accessed: 2025-02-11.

Zhou, Z., Rahman Siddiquee, M. M., Tajbakhsh, N., and Liang, J. (2018). Unet++: A nested u-net architecture for medical image segmentation. In Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support: 4th International Workshop, DLMIA 2018, and 8th International Workshop, ML-CDS 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 20, 2018, Proceedings 4, pages 3–11. Springer.

Zhu, Y., Hu, P., Li, X., Tian, Y., Bai, X., Liang, T., and Li, J. (2023). An end-to-end data-adaptive pancreas segmentation system with an image quality control toolbox. Journal of Healthcare Engineering, 2023(1):3617318.