Predicting Knee Osteoarthritis Progression Using DenseNet121 with Channel and Spatial Attention
Keywords:
Augmentation, Channel Attention Mechanism, Deep Learning, DenseNet121, Knee Osteoarthritis (OA), Mixup Data, MRI-based Image Classification, Osteoarthritis Prediction, Spatial Attention MechanismAbstract
The application of deep learning techniques for detecting and monitoring the progression of osteoarthritis (OA) is rapidly expanding. This study explores the predictive potential of MRI data combined with patient demographics and clinical information to forecast the onset and progression of knee OA. Specifically, the research focuses on predicting knee OA occurrence within a two-year timeframe by analyzing intermediate-weighted turbo spin-echo (IW-TSE) sequences from Osteoarthritis Initiative database.
We propose a novel methodology that integrates the DenseNet121 architecture with Mixup data augmentation and Channel and Spatial Attention mechanisms, aimed at improving image classification accuracy for knee OA incidence prediction. This approach addresses challenges associated with high intraclass variance and limited medical imaging datasets, by enhancing feature extraction and improving model generalization. We conducted experiments on a dataset comprising 186 MRI images across four classification categories, utilizing TensorFlow and Keras frameworks. The proposed methodology achieved a significant validation accuracy of 89.78%, demonstrating its effectiveness in predicting knee OA incidence.
These findings emphasize the potential of our methodology to enhance the accuracy of early-stage OA diagnosis and suggest a promising framework for medical image classification tasks. Moreover, the results provide a foundation for future research, optimizing deep learning models for clinical applications and advancing automated medical image analysis.
Downloads
References
D. J. Hunter, L. March, and M. Chew, “Osteoarthritis in 2020 and beyond: a Lancet Commission,” Lancet, vol. 396, no. 10264, pp. 1711–1712, 2020.
M. Cross et al., “The global burden of hip and knee osteoarthritis: estimates from the global burden of disease 2010 study,” Ann. Rheum. Dis., vol. 73, no. 7, pp. 1323–1330, 2014.
N. Arden and M. C. Nevitt, “Osteoarthritis: epidemiology,” Best Pract. Res. Clin. Rheumatol., vol. 20, no. 1, pp. 3–25, 2006.
D. T. Felson, “Osteoarthritis of the knee,” N. Engl. J. Med., vol. 354, no. 8, pp. 841–848, 2006.
S. Glyn-Jones, A. J. Palmer, R. Agricola, A. J. Price, T. L. Vincent, H. Weinans, and A. J. Carr, “Osteoarthritis,” Lancet, vol. 386, no. 9991, pp. 376–387, 2015.
Vashishtha and A. Kumar Acharya, “An overview of medical imaging techniques for knee osteoarthritis disease,” Biomed. Pharmacol. J., vol. 14, no. 2, pp. 903–919, 2021.
H. Zhang, M. Cisse, Y. Dauphin, and P. Jitkrittum, “Mixup: Beyond the empirical mixture of experts,” in International Conference on Machine Learning, 2017, pp. 788–797. https://arxiv.org/pdf/1710.09412
J. Hu, L. Shen, and G. Sun, “Squeeze-and-excitation networks,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018, pp. 7141–7151. https://arxiv.org/abs/2304.06502
V. Verma et al., "Manifold Mixup: Better Representations by Interpolating Manifolds," in International Conference on Learning Representations (ICLR), 2020.
X. Jin et al., "Automatic knee cartilage segmentation using 3D DenseNet-121 with attention mechanism," in Medical Imaging and Augmented Reality, 2019.
L. Yu et al., "Deep learning for knee osteoarthritis diagnosis: A review," Journal of Medical Imaging and Health Informatics.
X. Li et al., "DenseNet-121 for lung nodule classification in chest X-ray images," 2020.
L. Yu et al., "Employing DenseNet-121 for brain tumor segmentation in MRI scans," 2021.
J. Hu, L. Shen, and G. Sun, “Squeeze-and-Excitation Networks,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018, pp. 7132–7141.
L. Wang, Z. Q. Lin, and A. Wong, “COVID-Net: A Tailored Deep Convolutional Neural Network Design for Detection of COVID-19 Cases from Chest X-Ray Images,” Scientific Reports, vol. 10, no. 1, p. 19549, 2020.
X. Yu, J. Yang, N. Han, and K. L. Qin, “Deep Learning for Brain Tumor Classification,” in Brain Informatics, Springer, Cham, 2020, pp. 165–174.
S. Woo, J. Park, J.-Y. Lee, and I. So Kweon, “CBAM: Convolutional Block Attention Module,” in Proceedings of the European Conference on Computer Vision (ECCV), 2018, pp. 3–19.
Z. Li, Y. Wang, and J. Yu, “Attention-Based Deep Residual Learning Network for Liver Lesion Classification in CT Scans,” in IEEE Symposium on Computer-Based Medical Systems (CBMS), 2020, pp. 789–794.
M. Li et al., “A Deep Learning System for Automated Skin Lesion Diagnosis in Dermoscopic Images,” IEEE Journal of Biomedical and Health Informatics, vol. 25, no. 4, pp. 1239–1249, 2021.
H. Zhang, M. Cisse, Y. N. Dauphin, and D. Lopez-Paz, “Mixup: Beyond Empirical Risk Minimization,” in International Conference on Learning Representations, 2018. https://arxiv.org/pdf/1710.09412
J. Hu, L. Shen, and G. Sun, “Squeeze-and-Excitation Networks,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018, pp. 7132–7141. https://openaccess.thecvf.com/content_cvpr_2018/html/Hu_Squeeze-and-Excitation_Networks_CVPR_2018_paper.html
S. Woo, J. Park, J.-Y. Lee, and I. So Kweon, “CBAM: Convolutional Block Attention Module,” in Proceedings of the European Conference on Computer Vision (ECCV), 2018, pp. 3–19. https://openaccess.thecvf.com/content_ECCV_2018/html/Sanghyun_Woo_CBAM_Convolutional_Block_ECCV_2018_paper.html
L. Wang and A. Wong, “COVID-Net: A Tailored Deep Convolutional Neural Network Design for Detection of COVID-19 Cases from Chest X-Ray Images,” Scientific Reports, vol. 10, no. 1, p. 19549, 2020. https://www.nature.com/articles/s41598-020-76550-z
X. Yu, J. Yang, N. Han, and K. L. Qin, “Deep Learning for Brain Tumor Classification,” in Brain Informatics, vol. 7, 2020. https://link.springer.com/article/10.1007/s40708-020-00112-3
Srivastava, A., Hinton, G. E., Krizhevsky, A., & Dahl, G. E. (2014). Improving neural networks by preventing co-adaptation of feature detectors. arXiv preprint arXiv:1207.0580. Retrieved from https://arxiv.org/abs/1207.0580
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.
IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.
