Related Experiment Videos
A modified fuzzy C-means algorithm for bias field estimation and segmentation of MRI data
Mohamed N Ahmed1, Sameh M Yamany, Nevin Mohamed
1Systems and Biomedical Engineering Department, Cairo University, Giza, Egypt.
IEEE Transactions on Medical Imaging
|May 7, 2002
Summary
This study introduces a new fuzzy logic algorithm to improve magnetic resonance imaging (MRI) segmentation by addressing intensity inhomogeneities. The method enhances accuracy in segmenting MRI scans corrupted by noise.
Area of Science:
- Medical Imaging
- Computer Vision
- Fuzzy Logic
Background:
- Magnetic Resonance Imaging (MRI) data often suffers from intensity inhomogeneities.
- These artifacts, caused by hardware or acquisition issues, lead to errors in conventional intensity-based image classification.
- Existing methods struggle to accurately segment images with these shading artifacts.
Purpose of the Study:
- To develop a novel fuzzy logic algorithm for robust MRI segmentation.
- To estimate and compensate for intensity inhomogeneities in MRI data.
- To improve the accuracy of image classification in the presence of artifacts and noise.
Main Methods:
- A modified fuzzy c-means (FCM) algorithm is proposed.
- The objective function is altered to account for intensity inhomogeneities.
- A neighborhood influence mechanism is incorporated as a regularizer.
Main Results:
- The algorithm effectively compensates for MRI intensity inhomogeneities.
- Neighborhood regularization improves segmentation accuracy, particularly for noisy images.
- Experimental results demonstrate the algorithm's effectiveness on synthetic and real MR data.
Conclusions:
- The proposed fuzzy logic algorithm offers an effective solution for MRI segmentation with intensity inhomogeneities.
- The method enhances robustness against noise and artifacts, improving overall image analysis.
- This approach advances the field of medical image processing and analysis.