Diagnosing Knee Osteoarthritis Using Artificial Neural Networks and Deep Learning

Abstract

Knee osteoarthritis (OA) is a degenerative condition that significantly limits mobility and quality of life, particularly among older adults. Although radiographic assessment is the standard method for identifying structural changes, its accuracy can vary because interpretations depend heavily on the clinician\'s experience. Recent advances in artificial intelligence, intense learning, have enabled more objective and reliable analysis of medical images. In this work, we propose a hybrid diagnostic framework that combines a Convolutional Neural Network (CNN) to extract detailed radiographic features with an Artificial Neural Network (ANN) classifier to determine OA severity. The approach was trained and evaluated using publicly available knee X-ray datasets and demonstrated superior performance compared with traditional machine-learning techniques and standalone CNN models. Our findings indicate that integrating ANN-based decision layers with deep learning feature extractors can improve diagnostic consistency and assist healthcare professionals in detecting OA at earlier stages.

Introduction

Knee osteoarthritis (OA) is a chronic degenerative disease marked by cartilage breakdown, bone changes, and narrowing of joint space, leading to pain and reduced mobility. Although radiographs and the Kellgren–Lawrence (KL) grading scale are commonly used for diagnosis, these assessments are subjective and may vary between clinicians. Recent advances in artificial intelligence—especially deep learning—offer more objective tools for OA detection by automatically analyzing X-ray patterns.

This study proposes a hybrid CNN–ANN model to automate knee OA diagnosis and improve consistency. The CNN (a modified ResNet architecture) extracts key visual features such as joint-space width, osteophytes, bone texture, and structural abnormalities. These learned features are then fed into an ANN classifier with fully connected layers and dropout to refine decision-making and capture nonlinear relationships.

The model was trained on high-quality datasets including OAI and MOST. Preprocessing steps such as cropping, normalization, resizing, and augmentation improved image consistency and reduced overfitting. Training used the Adam optimizer, 50 epochs, and categorical cross-entropy loss.

Results show that the proposed CNN–ANN model significantly outperforms traditional methods:

• Accuracy: 89.8% (vs. 68.5% for SVM and 84.7% for ResNet-50)

• Strong gains in precision, recall, and F1-score

• Improved differentiation of mild, moderate, and severe OA, as shown by the confusion matrix

The discussion highlights that combining CNN-based feature extraction with ANN-based classification creates more robust decision boundaries, especially for overlapping categories. The model generalizes well across radiographs from multiple centers, making it suitable for clinical use. Future enhancements could integrate patient clinical data, improve model interpretability, and extend predictions to forecast OA progression.

Conclusion

This work introduces a hybrid deep learning model that integrates CNN-based feature extraction with an ANN classifier for the automated assessment of knee osteoarthritis on radiographic images. The proposed architecture demonstrated improved diagnostic performance when compared with traditional machine-learning methods and standalone CNN models. The results suggest that incorporating ANN-driven decision layers can enhance the reliability of OA grading and provide valuable support for clinicians, particularly in early disease identification. Future work can focus on integrating larger, multi-modal datasets to improve early detection and prediction of knee osteoarthritis. Enhancing model interpretability and real-time clinical deployment can further support automated decision-making for physicians.

References

[1] Antony, J., McGuinness, K., O’Connor, N. E., & Moran, K. (2017). Automatic knee osteoarthritis diagnosis from plain radiographs: A deep learning-based approach. arXiv. https://arxiv.org/abs/1710.10589 [2] Tiulpin, A., Thevenot, J., Rahtu, E., Lehenkari, P., &Saarakkala, S. (2019). Automatic grading of individual knee osteoarthritis features in plain radiographs using deep convolutional neural networks. Scientific Reports, 9, 1–10. https://doi.org/10.1038/s41598-019-45367-5 [3] Smith, A., & Kumar, R. (2025). CNN-ELM-based deep learning framework for knee osteoarthritis classification from radiographic images. International Journal of Advanced Science and Innovative Studies, 12(3), 45–60. [4] Liu, F., Zhou, Z., Jang, H., Samsonov, A., Zhao, G., Kijowski, R., & Bradshaw, T. (2022). Deep learning approach for early prediction of knee osteoarthritis using MRI. arXiv. https://arxiv.org/abs/2209.01192 [5] Zhang, Y., Zhao, X., & Li, H. (2024). Explainable AI for automated knee osteoarthritis detection: Grad-CAM and attention-based visualization. Journal of Medical Imaging, 11(2), 1–15. https://doi.org/10.1117/1.JMI.11.2.021203.

Copyright

Copyright © 2025 S. Bruntha, Dr. M. Chidambaram. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.