Education, Science, Technology, Innovation and Life
Open Access
Sign In
  1. Home
  2. Journals
  3. Advances in Computer, Signals and Systems
  4. Vol 9, Issue 4, 2025

A U-Net Baseline for Left Atrial Tumor Segmentation: Performance Analysis and Limitations

Download as PDF

DOI: 10.23977/acss.2025.090408 | Downloads: 0 | Views: 23

Author(s)

Yuhong Li 1

Affiliation(s)

1 Shenzhen Wisdom Nebula AI Technology Co., Ltd., Xili Street, Nanshan District, Shenzhen, China

Corresponding Author

Yuhong Li

ABSTRACT

Accurate and robust automated segmentation of Left Atrial (LA) tumors is essential for clinical diagnosis and treatment planning. Due to the inherent challenges in cardiac imaging, such as low tumor-to-background contrast and subtle boundaries, high-precision segmentation remains difficult. This study proposes and evaluates a standard 2D U-Net architecture for effective LA tumor segmentation. We address class imbalance using the Dice Loss function and enhance generalization through critical data augmentation, including elastic deformation. Evaluated on an independent cardiac MRI dataset, the U-Net model achieves a Dice Similarity Coefficient (DSC) of 0.8145, demonstrating its strong capability as a reliable baseline for this challenging task.

KEYWORDS

Left Atrial Tumor, Segmentation, U-Net, Cardiac MRI, Baseline Model

CITE THIS PAPER

Yuhong Li, A U-Net Baseline for Left Atrial Tumor Segmentation: Performance Analysis and Limitations. Advances in Computer, Signals and Systems (2025) Vol. 9: 62-68. DOI: http://dx.doi.org/10.23977/acss.2025.090408.

REFERENCES

[1] Ronneberger, O., Fischer, P., and Brox, T. (2015) U-Net: Convolutional Networks for Biomedical Image Segmentation. International Conference on Medical Image Computing and Computer-Assisted Intervention. Cham: Springer international publishing, 234-241.
[2] Milletari F., Navab N., and Ahmadi S.A. (2016) V-net: Fully convolutional neural networks for volumetric medical image segmentation.  The Fourth International Conference on 3D Vision (3DV). IEEE, 565-571.
[3] Oktay, O., Schlemper, J., Folgoc, L.L., Lee, M., Heinrich, M., Misawa, K., Mori, K., McDonagh, S., Jammerla, N.Y., Kainz, B., Glocker, B., and Rueckert, D. (2018) Attention u-net: Learning Where to Look for The Pancreas. arXiv preprint arXiv, 1804.03999.
[4] Jin, C., Feng, J., Wang, L., Yu, H., Liu, J., Lu, J., and Zhou, J. (2018) Left Atrial Appendage Segmentation and  Quantitative Assisted Diagnosis of Atrial Fibrillation Based on Fusion of Temporal-Spatial Information. Computers in biology and medicine, Vol. 96, 52-68.
[5] Li, J., Yu, Z. L., Gu, Z., Liu, H., and Li, Y. (2019) Dilated-Inception Net: Multi-Scale Feature Aggregation for Cardiac Right Ventricle Segmentation. IEEE Transactions on Biomedical Engineering, 66(12), 3499-3508.
[6] Sudre, C. H., Li, W., Vercauteren, T., Ourselin, S., and Jorge Cardoso, M. (2017) Generalised Dice Overlap as a Deep Learning Loss Function for Highly Unbalanced Segmentations. International Workshop on Deep Learning in Medical Image Analysis, 240-248. 
[7] Li, He. (2025) Left Atrium 2018, IEEE Dataport, doi:10.21227/5y8c-wf96.
[8] Jung, A. (2020) imgaug: Library for Image Augmentation in Machine Learning. Version 0.4.0. Available at: https://github.com/aleju/imgaug.

Subscription
E-Mail Alert
Most popular papers
Call for Special Issue
Publication Ethics & OA Statement
Plagiarism Policy
Contact Us
Downloads: 41294
Visits: 835820

Sponsors, Associates, and Links

All published work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright © 2016 - 2031 Clausius Scientific Press Inc. All Rights Reserved.