Related Experiment Video
Updated: Jun 4, 2025
02:09
Multi-modal Pulmonary Imaging: Using Complementary Information from CT and Hyperpolarized 129Xe MRI to Evaluate Lung Structure-Function
Published on: April 12, 2024
531
Robust thoracic CT image registration with environmental adaptability using dynamic Welsch's function and
Ziwen Wei1,2, Xiaolong Wu1, Ligang Xing3
1Anhui Province Key Laboratory of Medical Physics and Technology, Institute of Health and Medical Technology, Hefei Institutes of Physical Science, Hefei Cancer Hospital, Chinese Academy of Sciences, Hefei, China.
Quantitative Imaging in Medicine and Surgery
|December 19, 2024
Summary
This study introduces a new, robust algorithm for thoracic CT image registration, outperforming existing methods in noisy conditions. The approach enhances accuracy for clinical applications like surgical imaging.
Area of Science:
- Medical Imaging
- Computational Anatomy
- Image Registration
Background:
- Thoracic CT image registration is crucial but challenged by motion, high-density objects, and noise.
- Deep learning methods offer speed but often lack robustness.
- Existing methods struggle with noise, especially in low-dose CT scans.
Purpose of the Study:
- To develop a novel and robust algorithm for thoracic computed tomography (CT) image registration.
- To address limitations of current methods regarding noise and motion artifacts.
- To improve the reliability of thoracic CT image registration for clinical use.
Main Methods:
- Developed an anatomical structure-aware hierarchical registration approach.
- Utilized a divide-and-conquer strategy with region-specific dissimilarity metrics and regularization.
- Employed Welsch's function for flexible penalty distribution and a novel Welsch parameter update strategy for dynamic sparsity.
- Incorporated the majorization-minimization (MM) algorithm for efficient optimization.
Main Results:
- Achieved comparable performance to state-of-the-art methods in noise-free scenarios (TREs: 1.14-1.19 mm).
- Demonstrated superior robustness in the presence of noise, outperforming other methods (TREs: 1.78-2.38 mm).
- Ablation studies confirmed the effectiveness of individual components of the proposed method.
Conclusions:
- A novel and robust algorithm for thoracic CT image registration was successfully developed.
- The method shows significant potential for clinical applications, including surgical quantitative imaging.
- The algorithm effectively handles noise and motion, improving registration accuracy and reliability.
