Related Experiment Video
Updated: May 10, 2026
Multi-modal Pulmonary Imaging: Using Complementary Information from CT and Hyperpolarized 129Xe MRI to Evaluate Lung Structure-Function
Published on: April 12, 2024
Two phase non-rigid multi-modal image registration using Weber local descriptor-based similarity metrics and
Feng Yang1, Mingyue Ding, Xuming Zhang
1College of Life Science and Technology, Key Laboratory of Image Processing and Intelligent Control of Education Ministry of China, Huazhong University of Science and Technology, Wuhan 430074, China. fyang@foxmail.com
This study introduces a new two-phase method for non-rigid multi-modal medical image registration. Combining Weber Local Descriptors with normalized mutual information improves both accuracy and efficiency in aligning medical scans.
Area of Science:
- Medical Image Analysis
- Computer Vision
- Biomedical Engineering
Background:
- Non-rigid multi-modal image registration is crucial for medical image processing.
- Current methods using metrics like mutual information (MI) and sum of squared differences (SSD) lack both accuracy and efficiency.
Purpose of the Study:
- To develop a novel, efficient, and accurate two-phase non-rigid multi-modal image registration method.
- To address the limitations of existing registration techniques.
Main Methods:
- Proposed a two-phase registration approach combining Weber Local Descriptor (WLD) based similarity metrics with normalized mutual information (NMI).
- Employed the diffeomorphic free-form deformation (FFD) model.
- Phase 1: Recovered large deformations using WLD-based non-local SSD (wldNSSD) or WLD-based weighted structural similarity (wldWSSIM).
- Phase 2: Refined small deformations using NMI based on Phase 1 output.
Main Results:
- The proposed wldNSSD-NMI and wldWSSIM-NMI methods demonstrated superior performance compared to existing methods.
- Outperformed registration based on NMI, conditional mutual information (CMI), SSD on entropy images (ESSD), and ESSD-NMI.
- Achieved higher registration accuracy and computational efficiency in experiments.
Conclusions:
- The novel two-phase registration method significantly enhances accuracy and efficiency in non-rigid multi-modal medical image registration.
- The combination of WLD-based metrics and NMI offers a robust solution for complex image alignment tasks.
- This approach provides a valuable advancement for medical image analysis and processing applications.
