Related Experiment Video
Updated: May 15, 2026
04:48
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
Published on: July 5, 2024
Multi-organ segmentation with missing organs in abdominal CT images
Miyuki Suzuki1, Marius George Linguraru, Kazunori Okada
1Department of Computer Science, San Francisco State University, USA. miyukis@sfsu.edu
Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
|January 5, 2013
Summary
This study introduces novel methods for automatic missing organ detection (MOD) and multi-organ segmentation (MOS) in CT scans, improving accuracy for patients with surgically resected organs.
Area of Science:
- Medical Imaging
- Computer-Aided Diagnosis
- Surgical Pathology
Background:
- Multi-organ segmentation (MOS) in abdominal CT is crucial for diagnosis and treatment planning.
- Existing MOS methods struggle with cases involving missing organs due to surgical resection.
- Clinical populations frequently present with post-surgical anatomical variations.
Purpose of the Study:
- To develop an automatic missing organ detection (MOD) system for abdominal CT scans.
- To enhance state-of-the-art MOS to accurately segment abdominal organs in the presence of surgical resections.
- To improve the clinical applicability of automated segmentation for patients with altered anatomy.
Main Methods:
- Proposed automatic missing organ detection (MOD) by analyzing post-surgical organ motion and intensity homogeneity.
- Developed an atlas-based multi-organ segmentation (MOS) method capable of handling missing organs.
- Validated methods on 44 abdominal CT scans, including 9 cases with surgical organ resections.
Main Results:
- Achieved 93.3% accuracy in automatic missing organ detection (MOD).
- Demonstrated improved overall segmentation accuracy for the proposed MOS method on challenging diseased cases.
- Successfully segmented 10 abdominal organs in post-surgical patient data.
Conclusions:
- The proposed MOD and MOS methods effectively address the challenge of missing organs in abdominal CT.
- These advancements enable reliable segmentation for patients with post-surgical anatomical changes.
- The developed techniques hold significant potential for improving clinical workflow in radiology and surgery.
