Related Experiment Video
Updated: Dec 27, 2025
04:48
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
Published on: July 5, 2024
704
Block Level Skip Connections Across Cascaded V-Net for Multi-Organ Segmentation
IEEE Transactions on Medical Imaging
|February 25, 2020
Summary
This study introduces an efficient cascaded V-Net model with novel skip connections for improved multi-organ segmentation. The approach enhances segmentation accuracy, especially for challenging small organs, outperforming traditional methods.
Area of Science:
- Medical Image Analysis
- Computer Vision
- Machine Learning
Background:
- Multi-organ segmentation is complex due to label imbalance and organ variations.
- Accurate segmentation is crucial for medical diagnosis and treatment planning.
Purpose of the Study:
- To propose an efficient cascaded V-Net model for enhanced multi-organ segmentation.
- To address challenges in segmenting small, occluded, or boundary-unclear organs.
Main Methods:
- Developed a cascaded V-Net with dense Block Level Skip Connections (BLSC).
- Incorporated stacked small and large kernels with an inception-like structure.
- Implemented a two-stage approach for small organ segmentation: localization and single-class segmentation.
Main Results:
- Achieved average Dice score gains of 1.62% and 3.90% over traditional cascaded networks on SegTHOR and Multi-Atlas datasets.
- Improved segmentation for hard-to-segment small organs, with a 5.63% Dice score gain for the esophagus.
- Demonstrated significant performance improvements on the Multi-Atlas Labeling Beyond the Cranial Vault challenge with a 5.27% average Dice score gain for four organs.
Conclusions:
- The proposed cascaded V-Net with BLSC offers an efficient and effective solution for multi-organ segmentation.
- The novel approach significantly improves segmentation accuracy, particularly for challenging small organs.
- This method holds promise for advancing medical image analysis and clinical applications.
