Related Experiment Video
Updated: Aug 23, 2025
04:48
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
Published on: July 5, 2024
483
A combined deformable model and medical transformer algorithm for medical image segmentation.
Zhixian Tang1,2, Jintao Duan3, Yanming Sun3
1College of Medical Imaging, Shanghai University of Medicine & Health Sciences, Shanghai, 201318, China.
Medical & Biological Engineering & Computing
|November 3, 2022
Summary
This study introduces a novel approach combining deformable models and medical transformers to enhance medical image segmentation, particularly when training data is limited. The new method significantly improves segmentation accuracy for various medical imaging tasks.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Computer Vision
Background:
- Deep learning models for medical image segmentation often require large datasets, leading to generalization issues.
- Existing network structures can be inefficient, further limiting performance with scarce data.
Purpose of the Study:
- To develop a robust medical image segmentation method that overcomes data limitations.
- To improve the generalization capabilities of deep learning models in medical imaging.
Main Methods:
- A hybrid approach combining a statistical shape model for contour generation and thin plate splines for texture realism.
- Implementation of a medical transformer neural network for segmentation tasks.
Main Results:
- Achieved high segmentation accuracies: 89.97% for prostate MR images, 91.90% for heart US images, and 94.25% for tongue color images.
- Demonstrated improved performance in medical image segmentation compared to existing methods.
Conclusions:
- The proposed method effectively addresses the challenges of limited data in medical image segmentation.
- Combining deformable models with medical transformers offers a promising direction for enhancing segmentation accuracy and generalization.
