Abdominal CT Deformable Image Registration (DIR) Validation Dataset
收藏资源简介:
该数据集是首个用于腹部CT图像可变形配准(DIR)验证的基准数据集,由杜克大学和华盛顿大学医学院的研究团队创建。数据集包含30名患者的腹部CT图像对,每对图像包含63个平均的血管分叉点标记对,总计1895个标记对。数据来源于多个公开数据库及研究机构内部,经过深度学习模型分割和手动标记处理,确保标记对的高精度(0.7mm +/- 1.2 mm)。该数据集旨在支持DIR算法的开发与验证,解决腹部CT图像配准中的复杂变形问题,提升临床应用的精度与可靠性。
This is the first benchmark dataset for validating Deformable Image Registration (DIR) of abdominal CT images, developed by a research team from Duke University and the University of Washington School of Medicine. The dataset consists of 30 pairs of abdominal CT images from 30 patients, with each image pair containing 63 pairs of averaged vascular bifurcation landmarks, totaling 1895 landmark pairs. The dataset is sourced from multiple public databases and internal institutional research datasets, and has undergone deep learning model-based segmentation and manual landmark annotation to ensure high precision of the landmark pairs, with an error range of 0.7mm ± 1.2mm. This dataset is designed to support the development and validation of DIR algorithms, tackle the complex deformation challenges in abdominal CT image registration, and enhance the accuracy and reliability of clinical applications.

- 1A Vessel Bifurcation Landmark Pair Dataset for Abdominal CT Deformable Image Registration (DIR) Validation杜克大学, 华盛顿大学医学院 · 2025年



