Related Experiment Video
Updated: Jun 13, 2026
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
Effective fuzzy c-means based kernel function in segmenting medical images.
S R Kannan1, S Ramathilagam, A Sathya
1Department of Electrical Engineering, National Cheng Kung University, Tainan 70701, Taiwan. s.r.kannan@ruraluniv.ac.in
Computers in Biology and Medicine
|May 7, 2010
Summary
This study introduces a robust fuzzy c-means algorithm for segmenting noisy breast and brain MRI scans. The novel method improves clustering accuracy by using a kernel-induced distance measure and specialized initialization.
Area of Science:
- Medical Imaging
- Computer Vision
- Image Processing
Background:
- Conventional fuzzy c-means (FCM) struggles with medical images corrupted by noise and artifacts.
- The squared-norm distance in standard FCM is sensitive to outliers, limiting segmentation accuracy.
Purpose of the Study:
- Develop an effective robust fuzzy c-means (RFCM) algorithm for segmenting breast and brain MRI.
- Enhance medical image segmentation by addressing limitations of conventional FCM.
Main Methods:
- A novel objective function incorporating a robust kernel-induced distance measure was developed.
- Optimized equations for cluster centers and membership grades were derived.
- A specialized center initialization method was introduced to improve clustering performance.
Main Results:
- The proposed RFCM algorithm demonstrated superior performance in segmenting synthetic and real noisy medical images.
- Silhouette method validation confirmed the improved clustering validity.
- Comparative analysis showed the proposed method outperformed other recent FCM techniques.
Conclusions:
- The developed robust fuzzy c-means algorithm effectively segments noisy breast and brain MRI scans.
- The kernel-induced distance and specialized initialization significantly improve clustering accuracy and robustness.