Fracture Detection Dataset: X-Ray Images

This comprehensive dataset consists of X-ray images categorized into two primary classes:

1. Fractured: Images showing bone fractures.
2. Not Fractured: Images of healthy bones without any fractures.

Designed for developing machine learning and deep learning models, the dataset is meticulously organized into training, testing, and validation folders, ensuring seamless integration into fracture detection workflows.

Dataset Structure

The dataset is divided into three main directories, each containing subfolders for the two classes (fractured and not_fractured):

• Train Folder: For model training.

  • Total Images: 4,480
  • Subfolder Distribution:
    • fractured/: 2,240 images
    • not_fractured/: 2,240 images

• Test Folder: For evaluating model performance.

  • Total Images: 720
  • Subfolder Distribution:
    • fractured/: 360 images
    • not_fractured/: 360 images

• Validation Folder: For tuning hyperparameters and validating accuracy during training.

  • Total Images: 720
  • Subfolder Distribution:
    • fractured/: 360 images
    • not_fractured/: 360 images

Key Dataset Details

• Total Images: 5,920 X-ray scans.
• Image Formats: Standard formats, including .jpg and .png.
• Image Source: Verified X-ray scans of human bones to maintain medical relevance.

Applications of the Fracture Detection Dataset

This dataset is ideal for advancing research and development in the following areas:

1. Medical Diagnostics: Powering AI models for accurate and efficient fracture detection in healthcare.
2. Deep Learning Development: Training convolutional neural networks (CNNs) for binary image classification tasks.
3. Educational Use: Providing a practical resource for students and researchers focused on medical imaging and AI.

How to Use the Dataset

1. Download: Access the dataset  similar platforms.
2. Preprocessing: Perform resizing, normalization, or data augmentation for optimized model training.
3. Model Development:
   • Use the train folder to train models.
   • Leverage the validation folder for fine-tuning hyperparameters.
   • Evaluate model performance using the test folder.
4. Evaluation Metrics: Analyze results using accuracy, precision, recall, F1-score, and ROC-AUC.