Related Experiment Video
Updated: May 10, 2026
04:48
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
Published on: July 5, 2024
Shortest-path constraints for 3D multiobject semiautomatic segmentation via clustering and Graph Cut
Summary
We introduce a new vicinity prior model for multiobject segmentation using graph-based shortest-path constraints. This method accurately segments structures with similar intensities and improves boundary precision in medical images.
Area of Science:
- Medical image analysis
- Computer vision
- Computational imaging
Background:
- Multiobject segmentation is challenging, especially for structures with identical intensity profiles.
- Existing methods may struggle with precise boundary placement and computational efficiency.
Purpose of the Study:
- To develop a novel vicinity prior model for improved multiobject semiautomatic segmentation.
- To enhance segmentation accuracy and boundary precision in medical imaging.
Main Methods:
- Derivation of shortest-path constraints from graph models of structure adjacency.
- Integration into a joint centroidal Voronoi image clustering and Graph Cut framework.
- Development of a piecewise-constant vicinity prior model with multi-level penalization.
Main Results:
- The vicinity prior enables correct segmentation of distinct structures with identical intensity profiles.
- Improved precision in segmentation boundary placement across synthetic, simulated, and real medical images.
- The clustering approach enhances segmentation efficiency (runtime, memory) without compromising quality.
Conclusions:
- The proposed vicinity prior model significantly advances multiobject segmentation accuracy and boundary definition.
- The joint clustering and segmentation framework offers a robust and efficient solution for medical image analysis.
- This approach provides a controllable trade-off between boundary adaptivity and cluster compactness.
