🦴 Bone Fracture X-ray Dataset: Simple vs. Comminuted Fractures 🦴

🚀 Dive into a comprehensive dataset of X-ray images specifically curated for bone fracture classification, distinguishing between simple and comminuted fractures! This dataset is an invaluable resource for researchers and developers aiming to build and evaluate machine learning models in medical imaging.

🧑‍🔬 Contributors: Fahim Faisal Talha Talha, Md Soroar Jahan, Mayen Uddin Mojumdar
📅 Published: December 3, 2024
Version: v3
🔗 DOI: 10.17632/vg95gvhj3y.3

🔍 Dataset Overview

This dataset is meticulously organized to facilitate the development of robust models for image classification, segmentation, and fracture type recognition. It comprises high-quality X-ray images along with a substantial set of augmented images, ensuring a diverse and challenging training environment.

🦴 Fracture Categories

1. Simple Fracture:

   • Source: Exclusively sourced from hospital records for high clinical relevance.
   • Original Images: 1,211 images.
   • Augmented Images: 6,311 images.
   • Total Images: 7,522 images.

2. Comminuted Fracture:

   • Source: A blend of hospital-sourced images and web-sourced images (approximately one-tenth from web pages) to represent a broader range of cases.
   • Original Images: 1,173 images.
   • Augmented Images: 7,366 images.
   • Total Images: 8,539 images.

✨ Key Features

• Number of Original Images: 2,384
• Number of Augmented Images: 13,677
• Total Dataset Size: 16,061 images (Original + Augmented)
• File Format: JPG

🔄 Augmentation Techniques

To enhance the diversity and robustness of the dataset, the following augmentation techniques were applied:

• Zoom: Randomized scaling to simulate different viewing distances.
• Rotation: Rotations up to ±30° to account for variations in image orientation.
• Brightness and Contrast Adjustments: Random adjustments within the 80%–120% range to mimic varying lighting conditions.
• Flips: Both vertical and horizontal flips to introduce symmetry variations.

🗓️ Update Details (Version 3)

• Original Submission Date: November 15, 2024
• Update Date: December 2, 2024

Changes in This Version:

• Expanded Dataset Size: Increased significantly from 6,798 to 16,061 total images.
• Augmented Images:
  • Simple Fracture: Increased from 3,280 to 6,311 images.
  • Comminuted Fracture: Increased from 2,900 to 7,366 images.
• Augmentation Enhancements: Additional transformations were implemented to ensure a wider variety of realistic training images.

🎯 Applications

This dataset holds immense potential for various applications, including:

• Medical Imaging: Train and evaluate advanced models for accurate fracture classification and identification.
• Healthcare Technology: Support the development of innovative diagnostic tools and mobile applications for real-time fracture detection, potentially improving patient care.
• Medical Research: Facilitate a deeper understanding of fracture patterns and their visual characteristics.

🏥 Dataset Collection

• Simple Fracture Images: Sourced directly from hospitals to guarantee clinical accuracy and relevance.
• Comminuted Fracture Images: Obtained through a combination of hospital records and diverse web sources to represent a wide spectrum of imaging scenarios.

💡 Realistic Scenarios

The dataset is designed to simulate real-world clinical environments, featuring variations in lighting conditions, image orientations, and overall imaging environments. This approach ensures that models trained on this data are robust and generalizable.

📜 Citation

Talha, Fahim Faisal Talha; Jahan, Md Soroar; Mojumdar, Mayen Uddin (2024), “Bone Fracture X-ray Dataset: Simple vs. Comminuted Fractures”, Mendeley Data, V3, doi: 10.17632/vg95gvhj3y.3

🙏 If you find this dataset valuable for your research, please upvote! 👍 Your support is greatly appreciated! 😊