Integrating VGG Re-trained Feature Extraction with Machine Learning for Knee Osteoarthritis Severity Levels Detection Using X-Ray Images

Aug 4, 2026

Knee osteoarthritis is a degenerative joint disease whose diagnosis remains challenging because its progression and imaging interpretation can be subjective. The study develops an approach that combines fine-tuned VGG16 and VGG19 feature extraction with classical machine learning classifiers. A severity-graded Knee Osteoarthritis dataset is preprocessed, features are extracted by the retrained VGG models, and classifiers including Naive Bayes, k-NN, Decision Tree, Random Forest, Bagging, and AdaBoost are trained and evaluated. Results show VGG19 fine-tuning with Random Forest achieves the best performance, improving early detection accuracy and supporting clinical decision-making and personalized management.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages7

Reading Time18 min

AccessFree to download

Introduction

Osteoarthritis and knee osteoarthritis background

Diagnostic challenges and imaging limitations

Machine learning and deep learning for KOA detection

Related Approaches

Severity classification with traditional ML

CNN-based medical image analysis

ISSN ONLINE: 2447-0228 ITEGAM-JETIA Manaus, v.11 n.53, p. 36-42. May/June., 2025. DOI: [https://doi.org/10.5935/jetia. v11i53.1252](https://doi.org/10.5935/jetia. v11i53.1252) | RESEARCH ARTICLE | | | OPEN ACCESS | | --- | --- | --- | --- | | INTEGRATING VGG RE-TRAINED FEATURE EXTRACTION WITH MACHINE LEARNING FOR KNEE OSTEOARTHRITIS SEVERITY LEVELS
DETECTION USING X-RAY IMAGES
Simeon Yuda Prasetyo1 and Ghinaa Zain Nabiilah2
1,2Bina Nusantara University, Indonesia.
1[http://orcid.org/0000-0002-6077-4003](http://orcid.org/0000-0002-6077-4003), 2https://orcid.org/0000-0001-7638-7449
Email: [simeon.prasetyo@binus.ac.id](simeon.prasetyo@binus.ac.id),[ghinaa.nabiilah@binus.ac.id](ghinaa.nabiilah@binus.ac.id) | | | | | ARTICLE INFO | | ABSTRACT
Knee osteoarthritis, a degenerative joint disease affecting weight-bearing joints such as the knees and hips, poses substantial diagnostic hurdles due to its complicated pathophysiology and development. Traditional diagnostic methods rely heavily on clinical examinations and imaging techniques like X-rays, which can be subjective and vary with clinician experience. To overcome these problems, new advances in machine learning (ML) and deep learning (DL) offer promising alternatives for improving the accuracy of knee osteoarthritis identification. This study proposes a novel methodology that combines retrained VGG models with various machine learning techniques. The Knee Osteoarthritis Dataset with Severity Grading is preprocessed, and features are extracted using fine-tuned VGG16 and VGG19 models. A number of machine learning models, including Naive Bayes, K-Nearest Neighbors, Decision Tree, Random Forest, Bagging, and AdaBoost, are then trained using these extracted characteristics. These models' performance is assessed using metrics including F1-score, recall, accuracy, and precision. The results reveal that the combination of VGG19 with fine-tuning and Random Forest achieves the best performance, with an impressive accuracy of 62.68% . This approach significantly improves diagnostic accuracy and holds potential for enhancing clinical decision-making and management of knee osteoarthritis, offering a robust tool for early detection and personalized treatment strategies. | | | Article History
Received: August 16, 2024
Revised: January 20, 2025
Accepted: May 15, 2025
Published: May 31, 2025 | | | | | Keywords:
Knee osteoarthritis detection, Machine learning,
VGG Re-trained Feature Extraction,
X-ray imaging | | | | | | Copyright ©2025 by authors and Galileo Institute of Technology and Education of the Amazon (ITEGAM) . This work is licensed under the Creative Commons Attribution International License (CC BY 4.0) . | | | I. INTRODUCTION Weight-bearing joints, including the knees and hips, are susceptible to the complex illness known as osteoarthritis (OA) . Significantly contributing factors to its etiology include advanced age, high body mass index (BMI), and joint malalignment [1] . OAis a common type of arthritis that produces severe pain, stiffness, and swelling in the affected joints. Knee osteoarthritis in particular is one of the commonest forms of arthritis [2] . Knee osteoarthritis (KOA) is a slowly progressive disease that involves the degradation of cartilage, remodeling of bone, and inflammation [3] . The knee is the joint in the human body most commonly afflicted by this most prevalent musculoskeletal degenerative disease [4] . Pathologically, KOA is defined by a number of structural alterations in the knee joint, such as the development of osteophytes, inflammation of the synovium, subchondral sclerosis, and erosion of cartilage [5] . The impact of knee osteoarthritis extends beyond physical discomfort, as it is associated with a 35-37% increased risk of reduced time-to-mortality, primarily driven by pain [6] . Furthermore, KOA is linked to an increased risk of all-cause mortality, with disability and deterioration

FAQs

What problem does the study address in knee osteoarthritis diagnosis?

It addresses the difficulty of accurate knee osteoarthritis identification from X-ray images due to subjective interpretation and complex disease progression.

How does the proposed method combine VGG and machine learning?

The approach fine-tunes VGG16 and VGG19 to extract features, then trains multiple classifiers such as Naive Bayes, k-NN, Decision Tree, Random Forest, Bagging, and AdaBoost using those extracted features.

Which model combination achieves the best reported performance and why is it important?

VGG19 fine-tuning combined with Random Forest yields the best performance (best accuracy reported as 62.68%), supporting more accurate early detection and improved clinical decision-making for knee osteoarthritis severity.

Aug 4, 2026

Upload by River Wang