脑部影像自动配准数据集

本数据集供课题三元学习自推断医学影像诊断系统的脑部影像数据配准与融合部分使用
数据集内容:1250例T1脑部影像数据。
数据来源:来自于BraTs2021公开数据集。
采集地点与时间:2022年，网络下载。
采集方案：从BraTS2021公开数据集中下载原始影像数据，并对其进行标准化预处理，具体包括以下步骤：
数据格式转换：将原始影像数据统一转换为NIfTI格式，确保数据的兼容性和通用性。空间归一化：采用标准模板（如MNI152）对所有影像进行空间对齐，统一影像分辨率与体素大小。质量检查：检查影像的完整性和分辨率，对异常影像（如模糊、缺损）进行标记或剔除。数据标注：根据配准研究需求对影像进行标注，划分训练、验证和测试集，确保数据分布均匀。数据扩增：使用旋转、平移、弹性变形等方法进行数据扩增，模拟配准过程中的多样性。原始数据格式：以nii.gz文件格式存储，记录影像的三维结构信息，附带患者编号与标签。
设备状况：由于为公开数据集，数据采集过程未依赖具体设备，但数据质量经过BraTS官方审核，符合医学影像研究标准。

This dataset is designed for the registration and fusion of brain imaging data in the triplet learning-based self-inferential medical imaging diagnosis system project.
Dataset Content: 1250 cases of T1-weighted brain magnetic resonance imaging (MRI) data.
Data Source: Derived from the BraTS 2021 public dataset.
Collection Location and Time: Downloaded via public online sources in 2022.
Acquisition Protocol: The original imaging data was downloaded from the BraTS 2021 public dataset and underwent standardized preprocessing, including the following steps:
1. Data Format Conversion: Unify the original imaging data into NIfTI format to ensure data compatibility and generalizability.
2. Spatial Normalization: Align all imaging data to a standard template (e.g., MNI152) to unify the imaging resolution and voxel size.
3. Quality Inspection: Examine the integrity and resolution of the images, and mark or remove abnormal images such as blurry or defective ones.
4. Data Annotation: Annotate the images in accordance with the requirements of the registration research, and split them into training, validation, and test sets to ensure a uniform data distribution.
5. Data Augmentation: Apply data augmentation methods including rotation, translation, and elastic deformation to simulate the diversity encountered in the registration process.
Original Data Format: Stored in .nii.gz file format, which records the 3D structural information of the images, and is attached with patient identification numbers and labels.
Equipment Status: As a public dataset, the data collection process does not rely on specific medical imaging equipment. However, the data quality has been reviewed by the official BraTS committee and meets the standards of medical imaging research.

搜集汇总

背景与挑战

背景概述

该数据集专用于脑部影像数据的配准与融合研究，包含1250例源自BraTS2021公开数据集的T1脑部影像数据，并已进行标准化预处理。