Auto-ISC: Inter-Sample Contrastive Learning for Automated Curve Angle Calculation From Ultrasound

Chen Zhang , Anh Hoang Phuc Nguyen , Yongping Zheng , Sai Ho Ling

CAAI Transactions on Intelligence Technology ›› 2026, Vol. 11 ›› Issue (5) : 1321 -1332.

PDF (2684KB)
CAAI Transactions on Intelligence Technology ›› 2026, Vol. 11 ›› Issue (5) :1321 -1332. DOI: 10.1049/cit2.70157
ORIGINAL RESEARCH
research-article
Auto-ISC: Inter-Sample Contrastive Learning for Automated Curve Angle Calculation From Ultrasound
Author information +
History +
PDF (2684KB)

Abstract

Scoliosis assessment has been increasingly supported by AI-assisted tools, enabling more efficient and standardised clinical workflows. However, existing ultrasound-based pipelines often suffer from limited generalisation and unstable performance. These limitations primarily arise from two challenges: (1) suboptimal segmentation under low-contrast and noisy imaging conditions, and (2) unreliable anatomical landmark identification due to missing or ambiguous features. To address these issues, we propose Auto-ISC, a novel two-stage automated ultrasound scoliosis assessment system. The first stage employs the ISC-Swin segmentation model, enhanced by an Inter-Sample Contrastive Bank (ISCB) to improve cross-sample feature consistency and class-level discrimination. The second stage introduces a dynamic pairing logic algorithm for accurate matching of thoracic vertebrae, enabling reliable downstream angle estimation. Experimental results demonstrate strong segmentation performance (Dice: 86.10%, IoU: 76.08%) and high agreement with manual measurements. The proposed method achieves low average absolute differences (ABS) (thoracic: 2.83° ± 2.16°; lumbar: 2.89° ± 2.55°) and strong correlations (R2= 0.91 for thoracic and 0.81 for lumbar). Overall, Auto-ISC provides a reliable and radiation-free solution for automated scoliosis assessment, with improved clinical applicability.

Cite this article

Download citation ▾
Chen Zhang, Anh Hoang Phuc Nguyen, Yongping Zheng, Sai Ho Ling. Auto-ISC: Inter-Sample Contrastive Learning for Automated Curve Angle Calculation From Ultrasound. CAAI Transactions on Intelligence Technology, 2026, 11 (5) : 1321-1332 DOI:10.1049/cit2.70157

登录浏览全文

4963

注册一个新账户 忘记密码

References

[1]

C.-W. J. Cheung, G.-Q. Zhou, S.-Y. Law, T.-M. Mak, Ka-L. Lai, and Y.-P. Zheng, “ Ultrasound Volume Projection Imaging for Assessment of Scoliosis,” IEEE Transactions on Medical Imaging 34, no. 8 (2015): 1760-1768, https://doi.org/10.1109/tmi.2015.2390233.

[2]

H. Kim, H. S. Kim, E. S. Moon, et al., “ Scoliosis Imaging: What Radiologists Should Know,” Radiographics 30, no. 7 (2010): 1823-1842, https://doi.org/10.1148/rg.307105061.

[3]

Y.-P. Zheng, T. T.-Y. Lee, K. Ka-L. Lai, et al., “ A Reliability and Validity Study for Scolioscan: A Radiation-Free Scoliosis Assessment System Using 3d Ultrasound Imaging,” Scoliosis and Spinal Disorders 11 (2016): 1-15, https://doi.org/10.1186/s13013-016-0074-y.

[4]

J. Cobb, “ Outline for the Study of Scoliosis,” Instructional Course Lecture (1948).

[5]

A. R. Levy, M. S. Goldberg, N. E. Mayo, J. A. Hanley, and B. Poitras, “ Reducing the Lifetime Risk of Cancer From Spinal Radiographs Among People With Adolescent Idiopathic Scoliosis,” Spine 21, no. 13 (1996): 1540-1547, https://doi.org/10.1097/00007632-199607010-00011.

[6]

C. Huang, H. Tang, W. Fan, et al., “ Fully-Automated Analysis of Scoliosis From Spinal x-Ray Images,” in 2020 IEEE 33rd International Symposium on Computer-Based Medical Systems (CBMS) (IEEE, 2020), 114-119.

[7]

M. Zhao, N. Meng, J. P. Yin Cheung, C. Yu, P. Lu, and T. Z. Spinehrformer, “ A Transformer-Based Deep Learning Model for Automatic Spine Deformity Assessment With Prospective Validation,” Bioengineering 10, no. 11 (2023): 1333, https://doi.org/10.3390/bioengineering10111333.

[8]

Y. Cui, J. Zhu, Z. Duan, Z. Liao, S. Wang, and W. Liu, “ Artificial Intelligence in Spinal Imaging: Current Status and Future Directions,” International Journal of Environmental Research and Public Health 19, no. 18 (2022): 11708, https://doi.org/10.3390/ijerph191811708.

[9]

R. C. Brink, S. P. J. Wijdicks, I. N. Tromp, et al., “ A Reliability and Validity Study for Different Coronal Angles Using Ultrasound Imaging in Adolescent Idiopathic Scoliosis,” Spine Journal 18, no. 6 (2018): 979-985, https://doi.org/10.1016/j.spinee.2017.10.012.

[10]

J. Alison Noble and D. Boukerroui, “ Ultrasound Image Segmentation: A Survey,” IEEE Transactions on Medical Imaging 25, no. 8 (2006): 987-1010, https://doi.org/10.1109/tmi.2006.877092.

[11]

Y. Yao, W. Yu, Y. Gao, et al., “ W-Transformer: Accurate Cobb Angles Estimation by Using a Transformer-based Hybrid Structure,” Medical Physics 49, no. 5 (2022): 3246-3262, https://doi.org/10.1002/mp.15561.

[12]

Y. Tang, H. Chen, L. Qian, S. Ge, M. Zhang, and R. Zheng, “ Detection of Spine Curve and Vertebral Level on Ultrasound Images Using detr,” in 2022 IEEE International Ultrasonics Symposium (IUS) (IEEE, 2022), 1-4.

[13]

J. Lyu, S. Ho Ling, S. Banerjee, et al., “ Ultrasound Volume Projection Image Quality Selection by Ranking From Convolutional Ranknet,” Computerized Medical Imaging and Graphics 89 (2021): 101847, https://doi.org/10.1016/j.compmedimag.2020.101847.

[14]

C. Zhang, X. Deng, and S. Ho Ling, “ Next-Gen Medical Imaging: U-Net Evolution and the Rise of Transformers,” Sensors 24, no. 14 (2024): 4668, https://doi.org/10.3390/s24144668.

[15]

C. Zhang, Y. Zheng, J. McAviney, and S. Ho Ling, “ Ssat-Swin: Deep Learning-Based Spinal Ultrasound Feature Segmentation for Scoliosis Using Self-Supervised Swin Transformer,” Ultrasound in Medicine & Biology (2025).

[16]

S. Banerjee, J. Lyu, Z. Huang, et al., “ Light-Convolution Dense Selection u-Net (Lds u-Net) for Ultrasound Lateral Bony Feature Segmentation,” Applied Sciences 11, no. 21 (2021): 10180, https://doi.org/10.3390/app112110180.

[17]

Y. Zhou, Z. Huang, T. T.-Y. Lee, et al., “ Morphological-Consistent Diffusion Network for Ultrasound Coronal Image Enhancement,” arXiv preprint arXiv:2409.16661 (2024).

[18]

H.-D. Wu, W. Liu, and M.-S. Wong, “ Reliability and Validity of Lateral Curvature Assessments Using Clinical Ultrasound for the Patients With Scoliosis: A Systematic Review,” European Spine Journal 29, no. 4 (2020): 717-725, https://doi.org/10.1007/s00586-019-06280-y.

[19]

S. L. Weinstein, L. A. Dolan, J. C. Y. Cheng, A. Danielsson, and J. A. Morcuende, “ Adolescent Idiopathic Scoliosis,” lancet 371, no. 9623 (2008): 1527-1537, https://doi.org/10.1016/s0140-6736(08)60658-3.

[20]

C. Zhang, W. Jia, Y. Zheng, and S. Ho Ling, “ Isc-Swin: Inter Sample Contrastive Enhancement for Swin-Transformer in Ultrasound Spine Feature Segmentation,” in 2025 IEEE International Conference on Systems, Man, and Cybernetics (SMC) (IEEE, 2025), 1505-1510.

[21]

H. Xie, Z. Huang, F. H. F. Leung, et al., “ A Structure-Affinity Attention-Based Transformer Encoder for Spine Segmentation,” in 2024 IEEE International Symposium on Biomedical Imaging (ISBI) (IEEE, 2024), 1-5.

[22]

E. Xie, W. Wang, Z. Yu, A. Anandkumar, J. M. Alvarez, and P. Luo, “ SegFormer: Simple and Efficient Design for Semantic Segmentation With Transformers.” in Advances in Neural Information Processing Systems vol. 34 (2021), 12077-12090, https://doi.org/10.5555/3540261.3541185.

[23]

Hu Cao, Y. Wang, J. Chen, et al., “ Swin-Unet: Unet-Like Pure Transformer for Medical Image Segmentation,” in European Conference on Computer Vision (Springer, 2022), 205-218.

[24]

A. He, K. Wang, T. Li, C. Du, S. Xia, and H. Fu, “ H2former: An Efficient Hierarchical Hybrid Transformer for Medical Image Segmentation,” IEEE Transactions on Medical Imaging (2023).

[25]

J. Chen, Y. Lu, Q. Yu, et al., “ Transunet: Transformers Make Strong Encoders for Medical Image Segmentation,” arXiv preprint arXiv: 2102.04306 (2021).

[26]

R. Azad, M. Heidari, M. Shariatnia, et al., “ Transdeeplab: Convolution-Free Transformer-Based Deeplab v3+ for Medical Image Segmentation,” in International Workshop on PRedictive Intelligence In MEdicine (Springer, 2022), 91-102.

[27]

O. Ronneberger, P. Fischer, and T. Brox, “ U-Net: Convolutional Networks for Biomedical Image Segmentation,” in Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015: 18Th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III 18 (Springer, 2015), 234-241.

[28]

Z. Zhou, Md M. R. Siddiquee, N. Tajbakhsh, and J. Liang, “ 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 (Springer, 2018), 3-11.

[29]

A. Rykkje, J. F. Carlsen, and M. B. Nielsen, “ Hand-Held Ultrasound Devices Compared With High-end Ultrasound Systems: A Systematic Review,” Diagnostics 9, no. 2 (2019): 61, https://doi.org/10.3390/diagnostics9020061.

[30]

S.-Y. Ng and J. Bettany-Saltikov, “Imaging in the Diagnosis and Monitoring of Children With Idiopathic Scoliosis,” Open Orthopaedics Journal 11, no. 1 (2017): 1500-1520, https://doi.org/10.2174/1874325001711011500.

[31]

R. N. Uppot, “ Technical Challenges of Imaging & Image-Guided Interventions in Obese Patients,” British Journal of Radiology 91, no. 1089 (2018): 20170931, https://doi.org/10.1259/bjr.20170931.

[32]

M. J. Modica, K. M. Kanal, and M. L. Gunn, “ The Obese Emergency Patient: Imaging Challenges and Solutions,” Radiographics 31, no. 3 (2011): 811-823, https://doi.org/10.1148/rg.313105138.

[33]

A. Majeed, A. Abuhamad, L. Romary, and E. Sinkovskaya, “ Can Ultrasound in Early Gestation Improve Visualization of Fetal Cardiac Structures in Obese Pregnant Women?,” Journal of Ultrasound in Medicine 38, no. 8 (2019): 2057-2063, https://doi.org/10.1002/jum.14895.

[34]

R. Strony, “ Ultrasound-Assisted Lumbar Puncture in Obese Patients,” Critical Care Clinics 26, no. 4 (2010): 661-664, https://doi.org/10.1016/j.ccc.2010.07.002.

[35]

T. Ungi, F. King, M. Kempston, et al., “ Spinal Curvature Measurement by Tracked Ultrasound Snapshots,” Ultrasound in Medicine & Biology 40, no. 2 (2014): 447-454, https://doi.org/10.1016/j.ultrasmedbio.2013.09.021.

[36]

S. Banerjee, J. Lyu, Z. Huang, et al., “ Ultrasound Spine Image Segmentation Using Multi-Scale Feature Fusion Skip-Inception U-Net (SIU-Net),” Biocybernetics and Biomedical Engineering 42, no. 1 (2022): 341-361, https://doi.org/10.1016/j.bbe.2022.02.011.

[37]

Yi Huang, J. Jiao, J. Yu, Y. Zheng, and Y. Wang, “ Rsalunet: A Reinforcement Supervision U-Net-Based Framework for Multi-ROI Segmentation of Medical Images,” Biomedical Signal Processing and Control 84 (2023): 104743, https://doi.org/10.1016/j.bspc.2023.104743.

[38]

L. Cohen Monteiro, Y. Zheng, and S. Ho Ling, “ A Improved Hybrid Attention Recurrent Residual U-Net Approach for the Segmentation of Ultrasound Spine Images,” Biomedical Signal Processing and Control 108 (2025): 107925, https://doi.org/10.1016/j.bspc.2025.107925.

[39]

H. Xie, Z. Huang, Y. Zuo, et al., “ Sa2Net: Scale-Adaptive Structure-Affinity Transformation for Spine Segmentation From Ultrasound Volume Projection Imaging ,” Computerized Medical Imaging and Graphics 125 (2025): 102649, https://doi.org/10.1016/j.compmedimag.2025.102649.

[40]

Z. Huang, R. Zhao, F. H. F. Leung, et al., “ Joint Spine Segmentation and Noise Removal From Ultrasound Volume Projection Images With Selective Feature Sharing,” IEEE Transactions on Medical Imaging 41, no. 7 (2022): 1610-1624, https://doi.org/10.1109/tmi.2022.3143953.

[41]

S. Dana, C. Zhang, Y. Zheng, and S. Ho Ling, “ Vertebra-Level Completeness Analysis in Thoracolumbar Ultrasound Using a Yolo-Based Detection Framework,” Sensors (2026).

[42]

S. Banerjee, Z. Huang, J. Lyu, et al., “ Automatic Assessment of Ultrasound Curvature Angle for Scoliosis Detection Using 3-d Ultrasound Volume Projection Imaging,” Ultrasound in Medicine & Biology 50, no. 5 (2024): 647-660, https://doi.org/10.1016/j.ultrasmedbio.2023.12.015.

[43]

Y. Zhou, T. T.-Y. Lee, K. Ka-L. Lai, et al., “ Automatic Ultrasound Curve Angle Measurement via Affinity Clustering for Adolescent Idiopathic Scoliosis Evaluation,” Expert Systems with Applications 269 (2025): 126410, https://doi.org/10.1016/j.eswa.2025.126410.

[44]

Yi Huang, J. Jiao, J. Yu, Y. Zheng, and Y. Wang, “ Anatomy-Inspired Model for Critical Landmark Localization in 3d Spinal Ultrasound Volume Data,” Medical Image Analysis 103 (2025): 103610, https://doi.org/10.1016/j.media.2025.103610.

[45]

M. Gutmann and A. Hyvärinen, “ Noise-Contrastive Estimation: A New Estimation Principle for Unnormalized Statistical Models,” in Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics (JMLR Workshop and Conference Proceedings, 2010), 297-304.

[46]

I. Hacihaliloglu, “ Ultrasound Imaging and Segmentation of Bone Surfaces: A Review,” Technology 5, no. 2 (2017): 74-80, https://doi.org/10.1142/s2339547817300049.

PDF (2684KB)

0

Accesses

0

Citation

Detail

Sections
Recommended

/

〈 〉