Bone Fracture X-ray Dataset: Simple vs. Comminuted Fractures
收藏Mendeley Data2026-04-18 收录
下载链接:
https://data.mendeley.com/datasets/vg95gvhj3y
下载链接
链接失效反馈官方服务:
资源简介:
Dataset Overview:
This dataset has been curated to support research in bone fracture classification, focusing on simple and comminuted fractures. It includes high-quality X-ray images and a diverse set of augmented images to
facilitate the development and evaluation of machine learning models in medical imaging. The dataset is ideal for image classification, segmentation, and fracture type recognition tasks.
Fracture Categories:
1. Simple Fracture:
Source: Exclusively sourced from hospital records.
Original Images: 1,211 images.
Augmented Images: 6,311 images.
Total Images: 7,522 images.
2. Comminuted Fracture:
Source: A mix of hospital-sourced images and web-sourced images (approximately one-tenth from web pages).
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 Formats: JPG.
Augmentation Techniques:
Zoom: Randomized scaling.
Rotation: ±30° rotation.
Brightness and Contrast Adjustments: 80%–120% range.
Flips: Vertical and horizontal flips.
Update Details (Version 3):
Original Submission Date: 15 Nov 2024.
Update Date: 02 Dec 2024.
Changes in This Version:
Expanded Dataset Size:
From: 6,798 total images (previous version).
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 applied to ensure varied and realistic images for training.
Applications:
Medical Imaging:
Train and evaluate models for fracture classification and identification.
Healthcare Technology:
Support the development of diagnostic tools and mobile applications for real-time fracture detection.
Medical Research:
Aid in understanding fracture patterns and their visual indicators.
Dataset Collection:
Simple Fracture Images:
Exclusively sourced from hospitals to ensure clinical accuracy and relevance.
Comminuted Fracture Images:
Sourced from a mix of hospital records and web pages, representing diverse imaging conditions and perspectives.
Realistic Scenarios:
The dataset simulates real-world clinical settings with varying lighting conditions, orientations, and imaging environments to provide robust training data.
This dataset is a valuable resource for advancing medical imaging research and developing machine learning models tailored for fracture detection and classification tasks.
数据集概述:
本数据集专为骨折分类相关研究构建,重点覆盖单纯性骨折与粉碎性骨折两类场景。数据集包含高质量X线影像及多样化的增强影像,可助力医学成像领域机器学习模型的开发与评估工作,适配图像分类、图像分割及骨折类型识别等多项任务。
骨折分类类别:
1. 单纯性骨折:
数据来源:全部源自医院病历档案。
原始影像数量:1211张。
增强影像数量:6311张。
总影像量:7522张。
2. 粉碎性骨折:
数据来源:混合了医院档案影像与网络公开影像(约十分之一来自公开网页)。
原始影像数量:1173张。
增强影像数量:7366张。
总影像量:8539张。
核心参数:
原始影像总数量:2384张。
增强影像总数量:13677张。
数据集总规模:16061张影像(原始影像+增强影像)。
文件格式:JPG格式。
数据增强技术:
缩放:随机比例缩放。
旋转:±30°角度旋转。
亮度与对比度调整:调整范围为原值的80%至120%。
翻转:支持垂直翻转与水平翻转。
版本3更新详情:
首次提交日期:2024年11月15日。
更新日期:2024年12月2日。
本版本更新内容:
数据集规模扩容:原版本总影像量为6798张,本次更新后总影像量达16061张。
增强影像调整:
单纯性骨折增强影像:从3280张增至6311张。
粉碎性骨折增强影像:从2900张增至7366张。
增强流程优化:新增更多变换操作,确保训练用影像具备多样性与真实临床场景一致性。
应用场景:
医学成像领域:用于训练及评估骨折分类与识别相关模型。
医疗科技领域:助力开发实时骨折检测诊断工具及移动应用程序。
医学研究领域:辅助解析骨折形态特征及其视觉表征规律。
数据集采集方式:
单纯性骨折影像:全部源自医院档案,以保障临床准确性与场景相关性。
粉碎性骨折影像:混合采集自医院档案与公开网页,覆盖多样化的成像条件与拍摄视角。
真实场景模拟:
本数据集通过模拟不同光照条件、拍摄角度与成像环境,还原真实临床场景,可提供具备鲁棒性的训练数据。
本数据集可为医学成像研究的推进,以及适配骨折检测与分类任务的机器学习模型开发提供宝贵的研究资源。
创建时间:
2024-12-03



