Related Experiment Video
Updated: Jun 8, 2026
04:48
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
Published on: July 5, 2024
Multi-organ segmentation from multi-phase abdominal CT via 4D graphs using enhancement, shape and location
Marius George Linguraru1, John A Pura, Ananda S Chowdhury
1Imaging Biomarkers and Computer-Aided Diagnosis Laboratory, Radiology and Imaging Sciences, Clinical Center, National Institutes of Health, Bethesda, MD, USA. lingurarum@mail.nih.gov
Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
|October 1, 2010
Summary
This study presents an automated method for segmenting four abdominal organs from 4D CT scans using graph cuts. The approach optimizes computer-aided diagnosis (CAD) by incorporating anatomical and physiological information for improved accuracy.
Area of Science:
- Medical Imaging
- Computer-Aided Diagnosis
- Image Segmentation
Background:
- Accurate medical image interpretation requires anatomical and physiological priors for optimizing computer-aided diagnosis (CAD).
- Comprehensive analysis of multiple organs and quantitative soft tissue measures are crucial for diagnosis.
- Existing methods may lack efficiency in simultaneous multi-organ segmentation from dynamic CT data.
Purpose of the Study:
- To develop and evaluate an automated method for simultaneous segmentation of four abdominal organs from 4D CT data.
- To optimize computer-aided diagnosis (CAD) applications by leveraging anatomical and physiological priors.
- To investigate the impact of appearance, enhancement, shape, and location on organ segmentation accuracy.
Main Methods:
- Utilized graph cuts for simultaneous segmentation of four abdominal organs from 4D CT data.
- Employed contrast-enhanced CT scans (non-contrast and portal venous phases) with spatial normalization via non-linear registration.
- Incorporated 4D erosion with population historic information, CT enhancement, shape constraints (Parzen windows), and location priors (probabilistic atlas) into a 4D graph formulation.
Main Results:
- Demonstrated an automated method for simultaneous segmentation of liver, spleen, and kidneys from 4D CT data.
- Showcased the effectiveness of incorporating appearance, enhancement, shape, and location information for improved segmentation.
- Comparative results highlighted the influence of these factors on the accuracy of organ segmentation.
Conclusions:
- The developed automated 4D graph cut method enables simultaneous segmentation of multiple abdominal organs.
- Incorporating anatomical and physiological priors significantly enhances the accuracy of medical image segmentation for CAD.
- This approach holds promise for improving quantitative analysis and diagnostic capabilities in medical imaging.
