Exploiting Data Augmentation Strategies to Improve the Classification of Spinal Disorders in X-Ray Images

João Flávio Martins Pereira; João Fernando Mari; Leandro Henrique Furtado Pinto Silva

doi:10.22456/2175-2745.143521

Authors

João Flávio Martins Pereira Federal University of Viçosa (UFV) https://orcid.org/0009-0008-8511-9719
João Fernando Mari Federal University of Viçosa (UFV) https://orcid.org/0000-0001-8271-3802
Leandro Henrique Furtado Pinto Silva Federal University of Viçosa (UFV) https://orcid.org/0000-0002-5765-5206

DOI:

https://doi.org/10.22456/2175-2745.143521

Keywords:

Spinal Disorders, Deep Learning, Data Augmentation, CutMix, CutOut, MixUp

Abstract

Spinal disorders affect a significant portion of the population and are a growing concern for healthcare authorities due to their potentially debilitating consequences. In this context, computer vision techniques offer a promising path for rapid diagnosis, but they often require large datasets to train robust and reliable models. However, public datasets that can support the training of such models are scarce. To address this, we explored the application of advanced data augmentation strategies, namely CutMix, CutOut, and MixUp, combined with standard augmentation techniques to improve the performance of deep learning models in classifying X-ray images into three categories: (i) healthy, (ii) scoliosis, and (iii) spondylolisthesis. We applied these techniques by training the ResNet-50, Vision Transformer (ViT), and Swin Transformer V2 architectures, evaluating their effectiveness for this task. Our experiments revealed that the combination of ViT architecture and CutMix augmentation achieved the highest accuracy, with a performance of 0.9882.

Downloads

Download data is not yet available.

References

American Association of Neurological Surgeons. Scoliosis. 2021. https://www.aans.org/patients/conditions-treatments/scoliosis/. Accessed: August 22, 2024.

American Academy of Orthopaedic Surgeons. Spondylolysis and Spondylolisthesis. 2020. https://orthoinfo.aaos.org/en/diseases--conditions/spondylolysis-and-spondylolisthesis. Accessed: August 22, 2024.

Hospital de Cruzília. Adultos com dor na coluna. 2023. https://www.hospitaldecruzilia.com.br/site/index.php/noticias/todas-as-noticias/586-adultos-dor-na-coluna. Accessed: August 22, 2024.

KARPIEL, I. et al. A survey of methods and technologies used for diagnosis of scoliosis. Sensors, MDPI, v. 21, n. 24, p. 8410, 2021. Disponível em: https://doi.org/10.3390/s21248410.

CAO, C. et al. A survey of mix-based data augmentation: Taxonomy, methods, applications, and explainability. ACM Computing Surveys, ACM New York, NY, 2022. Disponível em: https://doi.org/10.1145/3696206.

TU, Y. et al. Automatic measurement algorithm of scoliosis Cobb angle based on deep learning. In: IOP PUBLISHING. Journal of Physics: Conference Series. 2019. v. 1187, n. 4, p. 042100. Disponível em: https://doi.org/10.1088/1742-6596/1187/4/042100.

HORNG, M.-H. et al. Cobb angle measurement of spine from X-ray images using convolutional neural network. Computational and Mathematical Methods in Medicine, Wiley Online Library, v. 2019, n. 1, p. 6357171, 2019. Disponível em: https://doi.org/10.1155/2019/6357171.

CHO, B. H. et al. Automated measurement of lumbar lordosis on radiographs using machine learning and computer vision. Global Spine Journal, SAGE Publications Sage CA: Los Angeles, CA, v. 10, n. 5, p. 611–618, 2020. Disponível em: https://doi.org/10.1177/2192568219868190.

FRAIWAN, M. et al. Using deep transfer learning to detect scoliosis and spondylolisthesis from X-ray images. PLOS ONE, Public Library of Science San Francisco, CA USA, v. 17, n. 5, p. e0267851, 2022. Disponível em: https://doi.org/10.1371/journal.pone.0267851.

ZHANG, T. et al. Deep learning model to classify and monitor idiopathic scoliosis in adolescents using a single smartphone photograph. JAMA Network Open, American Medical Association, v. 6, n. 8, p. e2330617–e2330617, 2023. Disponível em: https://doi.org/10.1001/jamanetworkopen.2023.30617.

WANG, H. et al. Application of deep learning upon spinal radiographs to predict progression in adolescent idiopathic scoliosis at first clinic visit. EClinicalMedicine, Elsevier, v. 42, 2021. Disponível em: https://doi.org/10.1016/j.eclinm.2021.101220.

CAMISA, A. et al. Automated detection of spinal lesions from CT scans via deep transfer learning. IEEE Access, IEEE, 2024. Disponível em: https://doi.org/10.1109/ACCESS.2024.3396999.

CHEN, Q. et al. Scoliosis detection with convolutional neural networks. In: IEEE. 2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM). 2022. p. 3785–3787. Disponível em: https://doi.org/10.1109/BIBM55620.2022.9995579.

SINGH, R. et al. Automated vertebrae diagnosis in spinal X-ray images using artificial intelligence. In: IEEE. 2023 4th International Conference for Emerging Technology (INCET). 2023. p. 1–6. Disponível em: https://doi.org/10.1109/INCET57972.2023.10170564.

LU, Q. T.; NGUYEN, T. M. An approach to classifying X-ray images of scoliosis and spondylolisthesis based on fine-tuned Xception model. International Journal of Advanced Computer Science & Applications, v. 15, n. 2, 2024. Disponível em: https://doi.org/10.14569/ijacsa.2024.0150262.

VERGARI, C.; SKALLI, W.; GAJNY, L. A convolutional neural network to detect scoliosis treatment in radiographs. International Journal of Computer Assisted Radiology and Surgery, Springer, v. 15, n. 6, p. 1069–1074, 2020. Disponível em: https://doi.org/10.1007/s11548-020-02173-4.

FRAIWAN, M.; AUDAT, Z.; MANASREH, T. A dataset of scoliosis, spondylolisthesis, and normal vertebrae X-ray images. Mendeley Data, 2022. Disponível em: https://doi.org/10.17632/xkt857dsxk.1.

HE, K. et al. Deep residual learning for image recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. [s.n.], 2016. p. 770–778. Disponível em: https://openaccess.thecvf.com/content_cvpr_2016/html/He_Deep_Residual_Learning_CVPR_2016_paper.html.

LIU, Z. et al. Swin transformer: Hierarchical vision transformer using shifted windows. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. [s.n.], 2021. p. 10012–10022. Disponível em: https://doi.org/10.48550/arXiv.2103.14030.

DOSOVITSKIY, A. et al. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. Disponível em: https://doi.org/10.48550/arXiv.2010.11929.

XU, M. et al. A comprehensive survey of image augmentation techniques for deep learning. Pattern Recognition, Elsevier, v. 137, p. 109347, 2023. Disponível em: https://doi.org/10.1016/j.patcog.2023.109347.

YUN, S. et al. Cutmix: Regularization strategy to train strong classifiers with localizable features. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. [s.n.], 2019. p. 6023–6032. Disponível em: https://doi.org/10.48550/arXiv.1905.04899.

DEVRIES, T.; TAYLOR, G. W. Improved regularization of convolutional neural networks with cutout. arXiv preprint arXiv:1708.04552, 2017. Disponível em: https://doi.org/10.48550/arXiv.1708.04552.

GUO, H.; MAO, Y.; ZHANG, R. Augmenting data with mixup for sentence classification: An empirical study. arXiv preprint arXiv:1905.08941, 2019. Disponível em: https://doi.org/10.48550/arXiv.1905.08941.

LOSHCHILOV, I.; HUTTER, F. SGDR: Stochastic gradient descent with warm restarts. In: International Conference on Learning Representations. [s.n.], 2022. Disponível em: https://doi.org/10.48550/arXiv.1608.03983.