A MULTI-SCALE ATTENTION U-NET MODEL FOR PANCREATIC TUMOR SEGMENTATION
Main Article Content
Abstract
Accurately segmenting pancreatic tumors from medical images is a major challenge in the field of computer vision due to the complex variations in tumor shape and size, and the low contrast between diseased tissue and adjacent organs. This study proposes an improved U-Net model to address the limitations of information loss and poor localization of traditional convolutional networks, enabling good detection of even small tumors. An attention mechanism is applied at the bypass connections to automatically weight key regions, allowing the model to focus on the tumor area and mitigate background noise. The proposed model, trained and evaluated on the Medical Segmentation Decathlon-Pancreas Task (MSD-PT) dataset, showed superior performance with a Dice coefficient (DSC) of 0.7 and a sensitivity of 0.76, confirming its effectiveness in supporting early detection and minimizing missed tumors.
Keywords
computed tomography, pancreatic tumor, tumor segmentation, U-Net model
Article Details
References
Bi, Y., Li, D., Pang, R., Du, C., Li, D., Zhao, X., & Lv, H. (2025). Diagnosis methods for pancreatic cancer with the technique of deep learning: A review and a meta-analysis. Frontiers in Oncology, 15, 1597969. https://doi.org/10.3389/fonc.2025.1597969
Bogovic, J. A., Prince, J. L., & Bazin, P.-L. (2013). A multiple object geometric deformable model for image segmentation. Computer Vision and Image Understanding, 117(2), 145-157. https://doi.org/10.1016/j.cviu.2012.10.006
Du, Y., Zuo, X., Liu, S., Cheng, D., Li, J., Sun, M., Zhao, X., Ding, H., & Hu, Y. (2023). Segmentation of pancreatic tumors based on multi-scale convolution and channel attention mechanism in the encoder-decoder scheme. Medical Physics, 50(1), 123-136. https://doi.org/10.1002/mp.16561
Huo, Y., Tang, Y., Chen, Y., Gao, D., Han, S., Bao, S., & Landman, B. A. (2019). Stochastic tissue window normalization of deep learning on computed tomography. Journal of Medical Imaging, 6(4), 044005-044005. https://doi.org/10.1117/1.JMI.6.4.044005
Jiang, H., Sun, Y., Liu, Y., & Wang, Y. (2020). DLU-Net: A dense learning U-Net for pancreas and pancreatic tumor segmentation on CT images. IEEE Access, 8, 221-232. https://doi.org/10.1109/BIBM49941.2020.9313263
Kaur, H., Kaur Saini, S., Thakur, N., & Juneja, M. (2024). Survey of denoising, segmentation and classification of pancreatic cancer imaging. Current Medical Imaging, 20, e150523216892. https://doi.org/10.2174/1573405620666230515090523
Li, Q., Liu, X., He, Y., Li, D., & Xue, J. (2023). Temperature guided network for 3D joint segmentation of the pancreas and tumors. Neural Networks, 157, 387–403. https://doi.org/10.1016/j.neunet.2022.10.026
Lu, L., Wu, M., Sen, G., Ren, F., & Hu, T. (2025). MDMU-Net: 3D multi-dimensional decoupled multi-scale U-Net for pancreatic cancer segmentation. PeerJ Computer Science, 11, e3059. https://doi.org/10.7717/peerj-cs.3059
Mahmoudi, T., Kouzahkanan, Z. M., Radmard, A. R., Kafieh, R., Salehnia, A., Davarpanah, A. H., Arabalibeik, H., & Ahmadian, A. (2021). Segmentation of Pancreatic Ductal Adenocarcinoma (PDAC) and surrounding vessels in CT images using deep convolutional neural networks and texture descriptors. bioRxiv. https://doi.org/10.1101/2021.06.09.447508
Pham, V., Nguyen, H., Pham, B., Nguyen, T., & Nguyen, H. (2023). Robust engineering-based unified biomedical imaging framework for liver tumor segmentation. Current Medical Imaging Reviews, 19(1), 37-45. https://doi.org/10.2174/1573405617666210804151024
Simpson, A. L., Antonelli, M., Bakas, S., Bilello, M., Farahani, K., van Ginneken, B., & Cardoso, M. J. (2019). A large annotated medical image dataset for the development and evaluation of segmentation algorithms. arXiv Preprint. https://doi.org/10.48550/arXiv.1902.09063
Sung, H., Ferlay, J., Siegel, R. L., Laversanne, M., Soerjomataram, I., Jemal, A., & Bray, F. (2021). Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: A cancer journal for clinicians, 71(3), 209-249. https://doi.org/10.3322/CAAC.21660
Tran, N. N., Nguyen, H. D., Huynh, N. T., Tran, N. P., & Nguyen, L. V. (2022). Segmentation on chest CT imaging in COVID-19 based on the improved attention U-Net model. In H. Fujita & M. Ali (Eds.), New trends in intelligent software methodologies, tools and techniques (pp. 596-606). IOS Press. https://doi.org/10.3233/FAIA220288
Yang, M., Zhang, Y., Chen, H., Wang, W., Ni, H., Chen, X., Li, Z., & Mao, C. (2022). AX-Unet: A deep learning framework for image segmentation to assist pancreatic tumor diagnosis. Frontiers in Oncology, 12, 894970. https://doi.org/10.3389/fonc.2022.894970
Zhang, Z., Keles, E., Durak, G., Taktak, Y., Susladkar, O., Gorade, V., Jha, D., Ormeci, A. C., Medetalibeyoglu, A., Yao, L., & Wang, B. (2024). Large-scale multi-center CT and MRI segmentation of pancreas with deep learning. Medical Image Analysis, 99, 103382. https://doi.org/10.1016/j.media.2024.103382
Zhang, Z., Yao, L., Keles, E., Velichko, Y., & Bagci, U. (2023). Deep learning algorithms for pancreas segmentation from radiology scans: A review. Advances in Clinical Radiology, 5(1), 31-52. https://doi.org/10.1016/j.yacr.2023.05.001