Related Experiment Video
Updated: May 25, 2026
14:08
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
Automated CT liver segmentation using improved Chan-Vese model with global shape constrained energy.
Xiuying Wang1, Chaojie Zheng, Changyang Li
1Biomedical and Multimedia Information Technology, Research Group, School of Information Technologies, University of Sydney, Australia. xiuying@it.usyd.edu.au
Summary
This study introduces an automated liver segmentation technique using a novel approach to address variations in liver shape and density. The method achieves accurate and robust liver segmentation in computed tomography (CT) images.
Area of Science:
- Medical Imaging
- Computer-Aided Diagnosis
- Biomedical Engineering
Background:
- Accurate liver segmentation is crucial for medical diagnosis and treatment planning.
- Conventional methods struggle with variations in liver shape, size, and similar density distributions with surrounding tissues.
Purpose of the Study:
- To develop an automated liver segmentation method overcoming challenges in shape variability and density similarity.
- To enhance the performance of statistical shape models for liver segmentation.
Main Methods:
- Utilized signed distance function to eliminate the need for landmark correspondence in Principal Component Analysis (PCA).
- Improved the Chan-Vese model to integrate shape energy and local intensity features for robust surface evolution.
- Employed PCA for shape-driven evolution, ensuring global and local accuracy.
Main Results:
- The proposed method demonstrated accurate and robust liver segmentation capabilities.
- Effective performance was observed in both low-contrast and high-contrast computed tomography (CT) images.
- Validation was performed on 25 clinical CT studies after training on 20.
Conclusions:
- The developed automated method offers a reliable solution for liver segmentation in CT imaging.
- The integration of signed distance functions and improved Chan-Vese model enhances segmentation accuracy and robustness.
- This technique holds potential for improving clinical workflows in liver disease diagnosis and management.
