Related Experiment Video
Updated: May 2, 2026
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
3.6K
Gamma-convergence approximation to piecewise smooth medical image segmentation.
Jungha An1, Mikael Rousson, Chenyang Xu
1Institute for Mathematics and its Applications (IMA), University of Minnesota, Minneapolis, MN, USA.
Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
|November 30, 2007
Summary
This study introduces a new variational region-based algorithm for accurate medical image segmentation, particularly for magnetic resonance (MR) images. The method effectively handles intensity variations, improving structure extraction in liver MR scans.
Area of Science:
- Medical Imaging
- Image Processing
- Computational Anatomy
Background:
- Accurate extraction of anatomical structures is challenging in medical imaging, especially with magnetic resonance (MR) images affected by acquisition protocols.
- Existing segmentation methods often struggle with spatial intensity perturbations and limitations of global or local region-based approaches.
Purpose of the Study:
- To develop a robust variational region-based algorithm for improved medical image segmentation.
- To address the challenges posed by spatial intensity variations and acquisition protocol dependencies in MR imaging.
Main Methods:
- A variational region-based algorithm utilizing gamma-Convergence approximation for a multi-scale piecewise smooth model.
- Implementation through efficient recursive Gaussian convolutions.
- Application to 2-dimensional human liver MR images.
Main Results:
- The proposed model effectively handles spatial perturbations in image intensity.
- It overcomes limitations of global region models and avoids high sensitivity of local approaches.
- Numerical experiments demonstrate favorable comparison with existing segmentation methods.
Conclusions:
- The developed algorithm offers a significant improvement for medical image segmentation, particularly for MR images.
- The method provides accurate structure extraction by effectively managing image quality variations.
- This approach shows promise for enhancing the analysis of medical imaging data.
