Related Experiment Video
Updated: May 3, 2026
04:48
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
Published on: July 5, 2024
996
Abdominal multi-organ CT segmentation using organ correlation graph and prediction-based shape and location priors
Toshiyuki Okada1, Marius George Linguraru2, Masatoshi Hori3
1Department of Radiology, Graduate School of Medicine Osaka University, 2-2 Yamadaoka, Suita, Osaka 565-0871, Japan. toshi@image.med.osaka-u.ac.jp
Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
|February 8, 2014
Summary
This study introduces a novel framework for automated multi-organ segmentation in upper abdominal CT scans. The method adapts to various imaging conditions without manual intensity data, achieving state-of-the-art performance.
Area of Science:
- Medical Imaging
- Computer Vision
- Radiology
Background:
- Automated organ segmentation in CT imaging is crucial for medical diagnosis and treatment planning.
- Existing methods often rely on manually traced training data, limiting adaptability to diverse imaging conditions.
Purpose of the Study:
- To develop an adaptable multi-organ segmentation framework for upper abdominal CT data.
- To eliminate the need for intensity information from manually traced training data.
Main Methods:
- Introduction of the organ correlation graph (OCG) to encode inherent spatial relationships between organs.
- Estimation of patient-specific organ shape and location priors using OCG.
- Generation of intensity priors from target and optionally untraced CT data.
Main Results:
- Successful segmentation of eight abdominal organs (liver, spleen, kidneys, pancreas, gallbladder, aorta, inferior vena cava).
- Evaluation on 86 CT datasets across four imaging conditions and two hospitals.
- Performance comparable to state-of-the-art methods using manually traced data.
Conclusions:
- The proposed framework offers an adaptable and effective solution for automated multi-organ segmentation in CT.
- The OCG-based approach reduces reliance on manual annotations, enhancing generalizability across imaging conditions.
