Related Experiment Video
Updated: Jul 20, 2026
06:48
Automated Segmentation of Cortical Grey Matter from T1-Weighted MRI Images
Published on: January 7, 2019
A knowledge-guided active model method of cortical structure segmentation on pediatric MR images
Zuyao Y Shan1, Carlos Parra, Qing Ji
1Division of Translational Imaging Research, Department of Radiological Sciences, St. Jude Children's Research Hospital, and Department of Biomedical Engineering, The University of Memphis, Tennessee 381005, USA. Zuyao.shan@stjude.org
Journal of Magnetic Resonance Imaging : JMRI
|August 25, 2006
Summary
A new automated method, the knowledge-guided active model (KAM), accurately quantifies pediatric brain structures on MR images. KAM shows superior agreement with manual segmentations compared to SPM2, aiding in treatment planning and disease evaluation.
Area of Science:
- Medical Imaging
- Computational Anatomy
- Pediatric Radiology
Background:
- Accurate segmentation of pediatric cortical structures is crucial for diagnosis and treatment.
- Existing automated methods may lack precision in complex pediatric brain anatomy.
Purpose of the Study:
- To develop an automated method for quantifying cortical structures on pediatric MR images.
- To evaluate the performance of the novel method against manual segmentation and existing software.
Main Methods:
- A knowledge-guided active model (KAM) utilizing Gibbs free energy principles was developed.
- Triangular mesh models were transformed and actively slithered to structure boundaries.
- KAM segmentation results were compared volumetrically and by similarity (kappa) to manual tracings and SPM2.
Main Results:
- KAM demonstrated higher average volumetric agreement (0.95) than SPM2 (0.90/0.80) with manual segmentations.
- KAM achieved superior similarity scores (kappa=0.95/0.93) compared to SPM2 (kappa=0.86).
- These findings held true for both healthy children and those with medulloblastoma.
Conclusions:
- A novel automated algorithm, KAM, was successfully developed for pediatric cortical structure segmentation on MR images.
- KAM shows better agreement with manual delineations than SPM2.
- KAM holds potential for applications in radiation therapy planning and quantitative disease assessment.
