Deep Learning Models for Brachial Lumen Segmentation: : A Comparative Study of UNet++, MultiResUNet, and TransUNet Architectures in B-Mode Ultrasound Imaging

Breno Luiz Nunes da Silva; Alexandre Gonçalves da Silva; Lucas Ziza Nastari; Rangel Arthur

doi:10.22456/2175-2745.150923

Authors

Breno Luiz Nunes da Silva Universidade Federal de Santa Catarina (UFSC) https://orcid.org/0009-0004-8287-2319
Alexandre Gonçalves Silva Universidade Federal de Santa Catarina (UFSC)
Lucas Ziza Nastari Universidade Estadual de Campinas (Unicamp)
Rangel Arthur Universidade Estadual de Campinas (Unicamp).

DOI:

https://doi.org/10.22456/2175-2745.150923

Keywords:

lumen segmentation, UNet++, MultiResUNet, TransUNet, deep learning

Abstract

Cardiovascular diseases remain one of the leading causes of global mortality. Flow-mediated dilation of the brachial artery, assessed by ultrasonography, is a non-invasive method to evaluate endothelial function; however, manual analysis is time-consuming and prone to variability. This study proposes the automatic segmentation of the brachial artery lumen using computer vision techniques based on deep learning. We investigated three segmentation architectures: UNet++, MultiResUNet, and TransUNet, applied to ultrasound images enhanced with contrast adjustment through CLAHE. Model training was conducted using K-fold cross-validation, with performance assessed through established metrics such as Dice, IoU, and Accuracy, while employing the Focal Tversky Loss function, suitable for imbalanced scenarios. Among the evaluated approaches, UNet++ achieved the best results, with Dice of 0.950, IoU of 0.905, and Accuracy of 0.980. These findings demonstrate the potential of deep segmentation models to reduce the subjectivity of clinical analysis and improve diagnostic accuracy in ultrasound-based vascular assessments.

Downloads

Download data is not yet available.

References

[1] World Health Organization. Cardiovascular Diseases (CVDs) — Fact Sheet. 2021. ⟨https://www.who.int/news-room/fact-sheets/detail/cardiovascular-diseases-(cvds⟩. Accessed: 2025-08-30.

[2] KASS, M.; WITKIN, A.; TERZOPOULOS, D. Snakes: Active contour models. International Journal of Computer Vision, Springer, v. 1, n. 4, p. 321–331, 1988.

[3] ZHOU, S.; CHELLAPPA, R.; MOGHADDAM, B. Tracking and segmenting multiple objects simultaneously using particle filters. IEEE Transactions on Pattern Analysis and Machine Intelligence, IEEE, v. 27, n. 3, p. 441–456, 2006.

[4] ZUIDERVELD, K. Contrast limited adaptive histogram equalization. In: Graphics Gems IV. [S.l.]: Academic Press, 1994. p. 474–485.

[5] RONNEBERGER, O.; FISCHER, P.; BROX, T. U-Net: Convolutional networks for biomedical image segmentation. Proceedings of MICCAI, Springer, p. 234–241, 2015.

[6] HUANG, Q. et al. Deep learning in medical ultrasound image analysis: A review. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, IEEE, v. 70, n. 1, p. 55–72, 2023.

[7] MARTINEZ, L.; PATEL, R.; WONG, K. Advances in vascular ultrasound imaging and segmentation: Emerging tools and clinical applications. Frontiers in Cardiovascular Medicine, Frontiers, v. 10, p. 1123456, 2023.

[8] SONG, Y. et al. Automated carotid plaque detection and segmentation in ultrasound using U-Net++. Computer Methods and Programs in Biomedicine, Elsevier, v. 214, p. 106549, 2022.

[9] ZHOU, L. et al. Automatic brachial artery segmentation in flow-mediated dilation ultrasound using deep learning and post-processing strategies. Ultrasound in Medicine & Biology, Elsevier, v. 49, n. 2, p. 345–356, 2023.

[10] AZZOPARDI, G.; CAMILLERI, K. et al. Vascular ultrasound image segmentation using convolutional neural networks: A comparative study. Biomedical Signal Processing and Control, Elsevier, v. 68, p. 102731, 2021.

[11] ZHANG, J. et al. Spatial attention multiscale network for carotid artery segmentation in ultrasound. Medical Image Analysis, Elsevier, v. 80, p. 102480, 2022.

[12] LI, W. et al. Hybrid CNN-Transformer architecture for peripheral artery ultrasound segmentation. Artificial Intelligence in Medicine, Elsevier, v. 152, p. 102435, 2024.

[13] NGUYEN, T.; ZHAO, M.; ROBERTS, P. Hybrid 3D deep learning architectures for ultrasound vascular segmentation. In: SPRINGER. Proceedings of MICCAI. [S.l.], 2023. p. 245–256.

[14] VASWANI, A. et al. Attention is all you need. Advances in Neural Information Processing Systems (NeurIPS), v. 30, p. 6000–6010, 2017.

[15] BUSLAEV, A. et al. Albumentations: Fast and flexible image augmentations. Information, MDPI, v. 11, n. 2, p. 125, 2020.

[16] BRADSKI, G. The OpenCV library. Dr. Dobb’s Journal of Software Tools, 2000.

[17] VINCENT, L. Morphological grayscale reconstruction in image analysis: Applications and efficient algorithms. IEEE Transactions on Image Processing, IEEE, v. 2, n. 2, p. 176–201, 1993.

[18] SILVA, A. G.; LOTUFO, R. de A. Efficient computation of new extinction values from extended component tree. Pattern Recognition Letters, v. 32, n. 1, p. 79–90, 2011. ISSN 0167-8655.

[19] VACHIER, C.; MEYER, F. Extinction value: A new measurement of persistence. In: Proceedings of IEEE Workshop on Nonlinear Signal and Image Processing. [S.l.: s.n.], 1995. p. 254–257.

[20] ZHOU, Z. et al. UNet++: A nested U-Net architecture for medical image segmentation. In: Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support. [S.l.]: Springer, 2018. (Lecture Notes in Computer Science, v. 11045), p. 3–11.

[21] IBTEHAZ, N.; RAHMAN, M. S. MultiResUNet: Rethinking the U-Net architecture for multimodal biomedical image segmentation. Neural Networks, Elsevier, v. 121, p. 74–87, 2020.

[22] CHEN, J. et al. TransUNet: Rethinking the U-Net architecture design for medical image segmentation through the lens of transformers. Medical Image Analysis, Elsevier, v. 92, p. 103280, 2024.

[23] CHEN, J. et al. TransUNet: Transformers make strong encoders for medical image segmentation. In: arXiv preprint arXiv:2102.04306. [S.l.: s.n.], 2021.

[24] ABRAHAM, N.; KHAN, N. M. A novel focal Tversky loss function with improved attention U-Net for lesion segmentation. Proceedings of ISBI, IEEE, p. 683–687, 2019.

[25] PEDREGOSA, F. et al. Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, v. 12, p. 2825–2830, 2011.

[26] LOSHCHILOV, I.; HUTTER, F. Decoupled weight decay regularization. International Conference on Learning Representations (ICLR), 2019.

[27] MICIKEVICIUS, P. et al. Mixed precision training. International Conference on Learning Representations (ICLR), 2018.

[28] PASZKE, A. et al. PyTorch: An imperative style, high-performance deep learning library. Advances in Neural Information Processing Systems (NeurIPS), v. 32, p. 8024–8035, 2019.

[29] SOILLE, P. Morphological Image Analysis: Principles and Applications. [S.l.]: Springer, 2003.

[30] TAHA, A. A.; HANBURY, A. Metrics for evaluating 3D medical image segmentation: analysis, selection, and tool. BMC Medical Imaging, BioMed Central, v. 15, n. 1, p. 29, 2015.

[31] SONG, Y. et al. Automatic carotid artery segmentation in ultrasound images using U-Net++. Computer Methods and Programs in Biomedicine, Elsevier, v. 221, p. 106899, 2022.