Related Experiment Video
Updated: Jun 8, 2026
10:25
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
Published on: September 25, 2019
Non-rigid registration with missing correspondences in preoperative and postresection brain images
Nicha Chitphakdithai1, James S Duncan
1Department of Biomedical Engineering, Yale University, New Haven, CT, USA. nicha.chitphakdithai@yale.edu
Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
|October 1, 2010
Summary
This study introduces a novel joint registration and segmentation algorithm to accurately align preoperative and postresection medical images, addressing the challenge of missing tissue data. The new method improves image alignment for better treatment evaluation.
Area of Science:
- Medical image analysis
- Computational anatomy
- Image registration
Background:
- Accurate registration of preoperative and postresection images is crucial for evaluating treatment effectiveness.
- Existing non-rigid registration methods struggle with datasets lacking tissue correspondence due to resections.
Purpose of the Study:
- To develop a joint registration and segmentation algorithm to address the missing correspondence problem in medical image alignment.
- To improve the accuracy of aligning preoperative and postresection images.
Main Methods:
- A novel algorithm combining registration and segmentation is presented.
- An intensity-based prior is utilized to segment the resection region.
- The Expectation-Maximization (EM) algorithm optimizes the maximum a posteriori (MAP) framework.
Main Results:
- The proposed method demonstrated improved image alignment compared to traditional non-rigid registration.
- Performance was superior to methods using robust error kernels in registration similarity metrics.
- Successful results were shown on both synthetic and real medical data.
Conclusions:
- The joint registration and segmentation approach effectively handles missing correspondences in medical imaging.
- This method offers enhanced accuracy for aligning pre- and post-resection images, aiding treatment assessment.
