Scholarlink

<p>Knee arthritis is a widespread musculoskeletal disorder that significantly impacts healthcare systems globally. Timely and accurate diagnosis is essential for improving patient outcomes and reducing the disease’s societal burden. This study presents a deep learning-based approach for the automatic detection and classification of knee arthritis into severe, moderate, and healthy categories using X-ray images. A dataset of 3000 knee X-rays was processed using advanced segmentation techniques, including Otsu binarization, noise removal, distance transformation, and watershed analysis, followed by contour-based feature extraction. Nine state-of-the-art convolutional neural network (CNN) models—DenseNet201, EfficientNetB0, EfficientNetB5, MobileNetV2, ResNet152V2, InceptionResNetV2, InceptionV3, Xception, and VGG19—were trained and evaluated. The analysis revealed that the VGG19 model achieved the highest accuracy when using the ADAM and RMSprop optimizers. Specifically, with ADAM, VGG19 achieved an impressive accuracy of 99.03% and a loss of 0.08, while with RMSprop, it achieved an accuracy of 97.16% and a loss of 0.28. Additionally, the training durations of the models were recorded. Among them, InceptionResNetV2 required the least training time, taking 1&#xa0;h and 15&#xa0;min, while Xception completed its training in just 1&#xa0;h using both ADAM and RMSprop optimizers respectively. The analysis of the work underscores the potential of optimized deep learning models in assisting clinicians with early diagnosis and treatment planning for knee arthritis. The research results contribute towards the development of robust, scalable, and generalizable AI-based diagnostic systems for clinical applications.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Optimized Deep Learning Techniques for Prediction and Prognosis of Severe and Moderate Knee Arthritis: A Comprehensive Analysis

  • 期刊论文

  • 发表日期：

    2025年9月4日

查看全文
  • Yogesh Kumar,
  • Pertik Garg,
  • Rupinder Singh,
  • Chamkaur Singh

摘要

Knee arthritis is a widespread musculoskeletal disorder that significantly impacts healthcare systems globally. Timely and accurate diagnosis is essential for improving patient outcomes and reducing the disease’s societal burden. This study presents a deep learning-based approach for the automatic detection and classification of knee arthritis into severe, moderate, and healthy categories using X-ray images. A dataset of 3000 knee X-rays was processed using advanced segmentation techniques, including Otsu binarization, noise removal, distance transformation, and watershed analysis, followed by contour-based feature extraction. Nine state-of-the-art convolutional neural network (CNN) models—DenseNet201, EfficientNetB0, EfficientNetB5, MobileNetV2, ResNet152V2, InceptionResNetV2, InceptionV3, Xception, and VGG19—were trained and evaluated. The analysis revealed that the VGG19 model achieved the highest accuracy when using the ADAM and RMSprop optimizers. Specifically, with ADAM, VGG19 achieved an impressive accuracy of 99.03% and a loss of 0.08, while with RMSprop, it achieved an accuracy of 97.16% and a loss of 0.28. Additionally, the training durations of the models were recorded. Among them, InceptionResNetV2 required the least training time, taking 1 h and 15 min, while Xception completed its training in just 1 h using both ADAM and RMSprop optimizers respectively. The analysis of the work underscores the potential of optimized deep learning models in assisting clinicians with early diagnosis and treatment planning for knee arthritis. The research results contribute towards the development of robust, scalable, and generalizable AI-based diagnostic systems for clinical applications.