Scholarlink

A distinctive approach is proposed in this paper that uses a convolutional neural network to automatically classify Knee Osteoarthritis (KOA) stages from X-ray images, aiming to create a powerful deep learning-based algorithm for KOA identification. To show how AI might enhance OA diagnosis and staging, the paper will examine the technology’s capacity to identify weak early-stage characteristics that are frequently overlooked by the human eye. The phases in the methodology are as follows: selecting a dataset, creating a Convolutional Neural Network (CNN) model, training it with appropriate loss functions and optimization techniques, and testing the model with metrics like accuracy, augmentation of data, interpretability strategies, and comparing it to previous approaches. Deep learning convolutional neural networks are being used in our model, VGG 16, RestNet150, EfficientNetB5, MobileNetV3 Small, and InceptionV3. EfficientNetB5 reigned supreme, achieving an impressive 90% test accuracy and proving its prowess for classifying our 5-classes image dataset.

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

Deep Learning-Based Classification of Knee Osteoarthritis Stages Using Convolution Neural Network on X-Ray Image

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Sajib Khan,
  • Najnin Sultana Nishu,
  • Mohammed Tasfiqur Rahman,
  • Saiful Islam Rumon,
  • Kamrul Islam,
  • Md. Samiul Islam,
  • Md Tanvir Chowdhury,
  • Habibur Rahman,
  • Monjurul Islam Sumon,
  • Eva Islam

摘要

A distinctive approach is proposed in this paper that uses a convolutional neural network to automatically classify Knee Osteoarthritis (KOA) stages from X-ray images, aiming to create a powerful deep learning-based algorithm for KOA identification. To show how AI might enhance OA diagnosis and staging, the paper will examine the technology’s capacity to identify weak early-stage characteristics that are frequently overlooked by the human eye. The phases in the methodology are as follows: selecting a dataset, creating a Convolutional Neural Network (CNN) model, training it with appropriate loss functions and optimization techniques, and testing the model with metrics like accuracy, augmentation of data, interpretability strategies, and comparing it to previous approaches. Deep learning convolutional neural networks are being used in our model, VGG 16, RestNet150, EfficientNetB5, MobileNetV3 Small, and InceptionV3. EfficientNetB5 reigned supreme, achieving an impressive 90% test accuracy and proving its prowess for classifying our 5-classes image dataset.