Advances in Knee Osteoarthritis Detection using Deep Learning and Content-Based Image Retrieval

Strategic Insights -

Researchers from Malatya Turgut Ozal University have developed a hybrid model to detect knee osteoarthritis using deep learning and content-based image retrieval (CBIR) techniques. The novel approach aims to overcome the limitations of existing deep learning models in distinguishing similar disease groups. The study highlights the challenges of increasing class diversity in deep learning architectures and proposes a hybrid model that leverages the benefits of different aspects of an image.

Key Takeaways:

  • The researchers used a developed hybrid CBIR-based system to detect knee osteoarthritis, which combines feature extraction using DarkNet53, Histogram of Oriented Gradients (HOG), and Local Binary Patterns (LBP) to leverage different aspects of an image.
  • The proposed model uses Neighborhood Component Analysis (NCA) to enhance speed and effectiveness, and seven different distance measurement metrics were used in the developed CBIR model.
  • Current deep learning architectures struggle to achieve comparable success rates in distinguishing between closely related but distinct disease groups.
  • The developed system aims to overcome the limitations of existing deep learning models in distinguishing similar disease groups.
  • The study highlights the challenges that increasing class diversity poses for the performance of deep learning architectures, especially when accessing large data sets.
  • CBIR techniques were preferred to detect knee osteoarthritis, and the proposed method uses a hybrid approach to combine the benefits of different image features.
  • The incidence of knee osteoarthritis increases, especially after a certain age, and the disease is a significant health concern worldwide.

Statistics:

  • 7 different distance measurement metrics were used in the developed CBIR model.
  • 3 different feature extraction methods were used: DarkNet53, Histogram of Oriented Gradients (HOG), and Local Binary Patterns (LBP).
  • 1 hybrid model was developed using the Neighborhood Component Analysis (NCA) method to enhance speed and effectiveness.
  • Knee osteoarthritis is a disease whose incidence increases, especially after a certain age.

Sources:

  • A novel approach in diagnosing knee osteoarthritis for content based image retrieval in big data analytics and medical images. Scientific Reports, 2025;15(1):29827.
  • Health & Medicine Week. August 29, 2025; p 845.
  • Nature Publishing Group. (www.nature.com/)
  • Scientific Reports. (www.nature.com/srep/)
  • Malatya Turgut Ozal University.