IRI-VFI dataset
- Citation Author(s):
-
Wei Si
(School of Statistics and Data Science, Nankai University)
Zhaolin Zheng (School of Statistics and Data Science, Nankai University)Zhewei Huang (stepfun, in China)Xi-Ming Xu (Big Data Engineering Center, Children's Hospital of Chongqing Medical University, National Clinical Research Center for Child Health and Disorders, Ministry of Education Key Laboratory of Child Development and Disorders, Chongqing 400016, China)Ruijue Wang (Hepatobiliary Surgery Children's Hospital of Chongqing Medical University, National Clinical Research Center for Child Health and Disorders, Ministry of Education Key Laboratory of Child Development and Disorders, Chongqing Key Laboratory of Pediatrics, C)Ji-Gang Bao (Tsinghua Shenzhen International Graduate School, Tsinghua University, Shenzhen, China and Shenzhen Wukong Investment Management Co. Ltd)Qiang Xiong (Hepatobiliary Surgery Children's Hospital of Chongqing Medical University, National Clinical Research Center for Child Health and Disorders, Ministry of Education Key Laboratory of Child Development and Disorders, Chongqing Key Laboratory of Pediatrics, C)Xiantong Zhen (the Central Research institute, United Imaging Healthcare Company Ltd., Shanghai 201815, China)Jun Xu (School of Statistics and Data Science, Nankai University) - Submitted by:
- Wei Si
- Last updated:
- DOI:
- 10.21227/v4kg-zp42
- Data Format:
226 views
- Categories:
- Keywords:
Abstract
We conducted a retrospective collection, covering 167 children who were examined and treated at the Children's Hospital of Chongqing Medical University from March 12, 2014 to January 7, 2022, with a total of 1634 IRI image sequences. This study has been registered with the Chinese Clinical Trial Registry, registration number ChiCTR2200058971, and complied with the provisions of the Declaration of Helsinki (DoH). The study was approved by the Institutional Ethical Review Board (document number 2022,69), and a waiver of informed consent was obtained.
Instructions:
Following data preprocessing, we obtained a final dataset comprising 1,357 sequences, totaling 95,308 images. These sequences encompass arterial, venous, and portal vein angiography imaging across multiple modalities such as Angiography, Fluoroscopy, and Subtraction.
We excluded abnormal IRI sequences exhibiting motion blur caused by the ARTIS icono ceiling or human motion, as well as static sequences devoid of blood flow. These sequences have the potential to distort the learning process of the blood flow interpolation model. Additionally, sequences lacking respiratory motion were excluded, as they offer limited information on vasomotion. Only sequences containing respiratory motion were retained.
Following data preprocessing, we obtained a final dataset comprising 1,357 sequences, totaling 95,308 images. These sequences encompass arterial, venous, and portal vein angiography imaging across multiple modalities such as Angiography, Fluoroscopy, and Subtraction.
We excluded abnormal IRI sequences exhibiting motion blur caused by the ARTIS icono ceiling or human motion, as well as static sequences devoid of blood flow. These sequences have the potential to distort the learning process of the blood flow interpolation model. Additionally, sequences lacking respiratory motion were excluded, as they offer limited information on vasomotion. Only sequences containing respiratory motion were retained.
Following data preprocessing, we obtained a final dataset comprising 1,357 sequences, totaling 95,308 images. These sequences encompass arterial, venous, and portal vein angiography imaging across multiple modalities such as Angiography, Fluoroscopy, and Subtraction.