|
|
 |
| |
|
|
|
Multi-Class Osteoporosis Detection Using Convolutional Neural Networks and Clinical Imaging Data |
|
|
|
PP: 721-731 |
|
|
doi:10.18576/amis/200312
|
|
|
|
Author(s) |
|
|
|
Arwa Darwish Alzughaibi,
Ebtesam Hussain Almansour,
Bashaer Almansour,
Tahani S. M. Shatir,
Tahani A. Esmail,
|
|
|
|
Abstract |
|
|
| Osteoporosis, Deep Learning, Convolutional Neural Networks (CNNs), Medical Image Processing, Bone Density Classification, Diagnostic Accuracy, Imaging Modalities (X-rays, CT, DEXA) and Healthcare Analytics. This study presents a deep learning-based approach employing Convolutional Neural Networks (CNNs) for the prediction and classification of osteoporosis. Leveraging multiple imaging modalities-namely X-rays, CT scans, and DEXA scans-the proposed framework integrates advanced medical data analysis with sophisticated image processing techniques. A custom CNN architecture is developed to categorize patients into three distinct groups: healthy, osteopenia, and osteoporosis. Extensive experimentation and comparisons with conventional machine learning techniques demonstrate that the CNN model outperforms traditional approaches in terms of accuracy, sensitivity, and specificity. The results consistently show improvements in classification performance across diverse datasets, achieving a test accuracy of 75 that offering valuable insights into feature relevance and model behavior. This work highlights the potential of deep learning to enhance diagnostic precision, facilitate early detection, and provide scalable, efficient solutions for osteoporosis management in clinical settings. |
|
|
|
|
|
|
|
