Related Experiment Video
Updated: May 21, 2026
10:25
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
Published on: September 25, 2019
Multi-Sectional Views Textural Based SVM for MS Lesion Segmentation in Multi-Channels MRIs.
Bassem A Abdullah1, Akmal A Younis, Nigel M John
1Department of Electrical and Computer Engineering, University of Miami, Miami, Fl, USA.
The Open Biomedical Engineering Journal
|June 29, 2012
Summary
This study introduces an automated method using textural features and support vector machines (SVM) for multiple sclerosis (MS) lesion segmentation in brain MRI. The approach enhances accuracy by analyzing multiple brain views, proving viable for clinical use.
Area of Science:
- Medical Imaging
- Machine Learning
- Neurology
Background:
- Multiple sclerosis (MS) lesion segmentation in brain MRI is crucial for diagnosis and monitoring.
- Current methods often require manual delineation, which is time-consuming and subjective.
Purpose of the Study:
- To develop a fully automated technique for segmenting MS lesions from brain MRI data.
- To improve the accuracy and efficiency of MS lesion detection using machine learning.
Main Methods:
- A novel approach employing a trained support vector machine (SVM) classifier.
- Utilizes textural features, along with other relevant features, for lesion discrimination.
- Employs multi-sectional view segmentation (axial, sagittal, coronal) for enhanced accuracy.
Main Results:
- The proposed textural-based SVM technique demonstrated effective MS lesion segmentation.
- Evaluation on simulated and real MRI datasets showed promising results compared to existing methods.
- The multi-sectional aggregation improved segmentation accuracy.
Conclusions:
- The developed automated method is a viable tool for clinical practice in detecting MS lesions.
- The use of textural features and SVM offers a robust approach to MS lesion segmentation.
- Multi-sectional analysis enhances the reliability of automated MRI segmentation for MS.
