Related Experiment Video
Updated: Aug 20, 2025
04:48
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
Published on: July 5, 2024
481
Auxiliary Segmentation Method of Osteosarcoma MRI Image Based on Transformer and U-Net
Feng Liu1,2, Jun Zhu3,4, Baolong Lv5
1School of Information Engineering, Shandong Youth University of Political Science, Jinan, Shandong, China.
Computational Intelligence and Neuroscience
|November 24, 2022
Summary
This study introduces OSTransnet, a novel method for segmenting osteosarcoma from MRI scans. It improves accuracy in identifying tumors, aiding clinical diagnosis and treatment.
Area of Science:
- Medical Image Analysis
- Oncology
- Artificial Intelligence
Background:
- Osteosarcoma is a prevalent malignant bone tumor with a poor prognosis.
- Manual identification of osteosarcoma in MRI is time-consuming and challenging due to image noise and blurred edges.
- Existing medical image processing methods struggle with the complex features of osteosarcoma MRI.
Purpose of the Study:
- To develop an automated osteosarcoma MRI image segmentation method (OSTransnet).
- To address challenges of fuzzy tumor edge segmentation and data noise-induced overfitting.
- To enhance the accuracy and stability of osteosarcoma lesion identification.
Main Methods:
- Dataset optimization including noise spatial distribution adjustment and data augmentation via image rotation.
- Segmentation using a hybrid U-Net and Transformer model with channel-based attention.
- Integration of an edge enhancement module (BAB) and a combined loss function.
Main Results:
- OSTransnet demonstrated accuracy and stability in segmenting osteosarcoma from over 4,000 MRI images.
- The method effectively improved fuzzy tumor edge segmentation and reduced overfitting.
- Validation confirmed the method's utility as an adjunct to clinical diagnosis.
Conclusions:
- OSTransnet offers a robust solution for osteosarcoma MRI segmentation.
- The proposed method enhances diagnostic capabilities by improving lesion identification accuracy.
- This AI-driven approach shows significant potential in clinical applications for osteosarcoma management.
