Related Experiment Video
Updated: Jul 9, 2026
06:45
Automated Joint Space Detection Improves Bone Segmentation Accuracy
Published on: November 28, 2025
Groupwise combined segmentation and registration for atlas construction
Kanwal K Bhatia1, Paul Aljabar, James P Boardman
1Visual Information Processing, Department of Computing, Imperial College London.
Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
|December 7, 2007
Summary
This study introduces new methods for creating average anatomical atlases, improving image alignment and tissue segmentation simultaneously. These novel atlases help quantify tissue growth in preterm infants between one and two years old.
Area of Science:
- Medical image analysis
- Computational anatomy
- Neuroimaging
Background:
- Average anatomical atlases are valuable for understanding typical anatomy and variations.
- Existing methods often focus on intensity-based atlases, neglecting detailed tissue segmentation.
- There is a need for robust methods to create atlases for populations lacking prior detailed anatomical data, such as preterm infants.
Purpose of the Study:
- To develop novel groupwise methods for simultaneously performing image registration and tissue segmentation.
- To create average 3D Magnetic Resonance (MR) atlases for preterm infants at one and two years of age.
- To utilize these atlases for quantifying tissue growth during this critical developmental period.
Main Methods:
- An iterative Expectation-Maximization (EM) framework was employed for simultaneous segmentation and registration.
- Groupwise approaches were used to align multiple images to a common average space.
- 3D MR images of preterm infants at one and two years old were used to build the atlases.
Main Results:
- Novel groupwise combined segmentation and registration approaches were successfully developed.
- Average 3D MR atlases were constructed for preterm infants at one and two years old.
- The developed atlases enabled the quantification of tissue growth between these two age points.
Conclusions:
- The presented methods effectively create average anatomical atlases with simultaneous segmentation and registration.
- These novel atlases provide a valuable tool for studying neurodevelopmental trajectories in preterm populations.
- The quantification of tissue growth highlights the utility of these atlases in developmental research.
