Related Experiment Video
Updated: Jun 13, 2026
14:08
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
Automatic model-guided segmentation of the human brain ventricular system from CT images
Jimin Liu1, Su Huang, Volkau Ihar
1Biomedical Imaging Lab, Singapore Bioimaging Consortium, 30 Bioplolis Street #07-01, Matrix, Singapore. liujm@sbic.a-star.edu.sg
Academic Radiology
|May 12, 2010
Summary
This study presents a model-guided method for automated brain ventricular system segmentation on CT scans. The approach achieves 85% overlap with expert segmentations, offering a fast and reproducible solution for neurodiagnosis.
Area of Science:
- Medical imaging analysis
- Computational neuroanatomy
Background:
- Accurate segmentation of the brain ventricular system is crucial for neurodiagnosis and neurosurgery.
- Manual segmentation is time-consuming, subjective, and lacks reproducibility.
- Existing automatic methods struggle with CT image noise and anatomical variations.
Purpose of the Study:
- To develop and evaluate a model-guided method for automated segmentation of the brain ventricular system on CT images.
- To address limitations of manual segmentation and current automatic techniques.
Main Methods:
- A five-step model-guided segmentation process was applied to 50 patient CT scans.
- Key steps include model registration, region specification, intensity thresholding for cerebrospinal fluid, individual ventricle segmentation, and calcification identification for refinement.
Main Results:
- The proposed method achieved an average overlap ratio of 85% compared to expert-derived ground truths for the entire ventricular system.
- Segmentation of each CT dataset required approximately 10 seconds on a standard desktop computer.
Conclusions:
- The model-guided method provides acceptable automated segmentation of the brain ventricular system, even with image noise and anatomical variations.
- This technique shows potential for quantitative interpretation of CT images in clinical neuroimaging settings.
