Related Experiment Video
Updated: May 13, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.6K
Human visual perception-inspired medical image segmentation network with multi-feature compression
Guangju Li1, Qinghua Huang2, Wei Wang3
1School of Computer Science, Northwestern Polytechnical University, Xi'an 710129, China; School of Artificial Intelligence, OPtics and ElectroNics (iOPEN), Northwestern Polytechnical University, Xi'an 710072, China.
Artificial Intelligence in Medicine
|April 25, 2025
Summary
This study introduces MS-Net, a novel medical image segmentation network inspired by human vision. It achieves state-of-the-art accuracy by effectively filtering noise and refining segmentation, outperforming existing methods with fewer parameters.
Area of Science:
- Medical image analysis
- Computer vision
- Neuroscience-inspired AI
Background:
- Medical image segmentation is vital for diagnosis and treatment planning.
- Current methods struggle with noise during feature fusion.
- Human visual system effectively suppresses noise and integrates features.
Purpose of the Study:
- To develop a medical image segmentation network inspired by human visual perception.
- To address the limitations of existing methods in handling noise during feature fusion.
- To improve segmentation accuracy and efficiency in medical imaging.
Main Methods:
- Proposed MS-Net, incorporating a multi-feature compression (MFC) module.
- MFC module mimics human visual processing to filter irrelevant features.
- Segmentation refinement (SR) module emulates physician-led lesion segmentation.
Main Results:
- MS-Net achieved state-of-the-art segmentation performance on three public datasets.
- Significantly reduced the number of parameters compared to existing models.
- Demonstrated effective noise suppression and precise boundary delineation.
Conclusions:
- MS-Net offers a novel, human vision-inspired approach to medical image segmentation.
- The network achieves superior accuracy and efficiency with reduced computational cost.
- This method holds promise for advancing computer-aided diagnosis and treatment planning.
