Related Experiment Video
Updated: Jul 6, 2026
04:48
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
Published on: July 5, 2024
Self-repelling snakes for topology-preserving segmentation models
Carole Le Guyader1, Luminita A Vese
1IRMAR, UMR CNRS 6625, Institut National des Sciences Appliquées de Rennes, 35043 Rennes Cedex, France. cleguyad@insa-rennes.fr
Summary
The level-set method tracks fronts but can alter topology. This study introduces a topology-preserving segmentation model for applications like medical imaging, ensuring shapes maintain their original structure.
Area of Science:
- Computer Vision
- Image Processing
- Computational Science
Background:
- The level-set method excels at tracking propagating fronts due to its Eulerian formulation and ability to handle topological changes.
- However, this topological flexibility is disadvantageous when preserving the initial shape's topology is crucial, as in medical image segmentation.
Purpose of the Study:
- To develop a novel image segmentation model that enforces topological constraints.
- To address the limitations of standard level-set methods in applications requiring topology preservation.
Main Methods:
- Utilizes an implicit level-set formulation.
- Incorporates geodesic active contours.
- Enforces a topological constraint within the segmentation process.
Main Results:
- The proposed model successfully segments images while preserving the topology of the evolving contour.
- Demonstrates applicability in scenarios where topological integrity is essential, such as human cortex reconstruction.
Conclusions:
- The developed topology-preserving level-set model offers a robust solution for image segmentation tasks with strict topological requirements.
- This approach enhances the reliability of segmentation in fields like medical imaging.