Related Experiment Video
Updated: Dec 17, 2025
10:25
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
Published on: September 25, 2019
49.1K
Groupwise registration with global-local graph shrinkage in atlas construction
Tianyu Fu1, Jian Yang2, Qin Li3
1School of Life Science, Beijing Institute of Technology, Beijing 100081, China; Beijing Engineering Research Center of Mixed Reality and Advanced Display, School of Optics and Photonics, Beijing Institute of Technology, Beijing 100081, China.
Medical Image Analysis
|June 26, 2020
Summary
This study introduces a novel global-local graph shrinkage method for accurate atlas construction. By integrating local image distributions, it enhances groupwise registration accuracy, especially for organs with high anatomical variability.
Area of Science:
- Medical Image Analysis
- Computational Anatomy
- Computer Vision
Background:
- Graph-based groupwise registration is standard for atlas construction.
- Conventional methods overlook local image distribution and structure, reducing accuracy for variable anatomy.
Purpose of the Study:
- To propose a global-local graph shrinkage approach for accurate atlas generation.
- To address limitations of conventional methods in handling inter-subject anatomical variability.
Main Methods:
- Construct a graph based on global image similarities.
- Incorporate local image distributions via clustering to simplify graph edges.
- Refine deformations using local similarities for warping along graph edges.
Main Results:
- The global-local graph shrinkage method was evaluated on synthetic and clinical liver datasets.
- Compared against six state-of-the-art methods, the proposed approach demonstrated superior accuracy.
- The method effectively respects both global and local features during atlas construction.
Conclusions:
- The proposed global-local graph shrinkage method significantly improves atlas generation accuracy.
- This approach is particularly beneficial for datasets with substantial inter-subject anatomical variations.
- It offers a more robust and accurate solution for population atlas construction.
