Related Experiment Video
Updated: Jun 26, 2026
14:08
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
[MR brain image segmentation based on modified fuzzy C-means clustering using fuzzy GIbbs random field]
Liang Liao1, Tusheng Lin, Bi Li
1School of Electronic and Information Engineering, South China University of Technology, Guangzhou 510640, China. liaoliangis@126.com
Summary
This study introduces a novel algorithm for segmenting noisy Magnetic Resonance (MR) brain images. The enhanced method improves accuracy by incorporating spatial constraints into fuzzy clustering, offering a better alternative for medical image analysis.
Area of Science:
- Medical Imaging
- Computer Vision
- Artificial Intelligence
Context:
- Magnetic Resonance (MR) imaging is crucial for brain analysis.
- Image segmentation is vital for extracting meaningful information from MR scans.
- Noise in MR images often degrades segmentation accuracy.
Purpose:
- To develop a modified algorithm for improved Magnetic Resonance (MR) brain image segmentation.
- To enhance the traditional fuzzy c-means (FCM) clustering algorithm by integrating spatial constraints.
- To address the challenge of noise in medical MR image segmentation.
Summary:
- A modified algorithm combines a fuzzy Gibbs random field model with fuzzy c-means (FCM) clustering for MR brain image segmentation.
- Spatial constraints, including homogeneity of cliques and fuzzy Gibbs clique potential, are introduced.
- A new objective function and iterative formulas enhance the traditional intensity-based FCM algorithm.
Impact:
- The proposed algorithm demonstrates improved performance in segmenting noisy MR brain images.
- Experiments on synthetic data and MR phantoms validate the algorithm's effectiveness.
- This method offers a superior alternative for segmenting medical MR images corrupted by noise.
