Related Experiment Video
Updated: Jan 4, 2026
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
Published on: September 25, 2019
Data on MRI brain lesion segmentation using K-means and Gaussian Mixture Model-Expectation Maximization
Ju Qiao1, Xuezhu Cai2, Qian Xiao3
1Department of Mechanical and Industrial, Northeastern University, Boston, MA, USA.
Abstract:
The data in this article provide details about MRI lesion segmentation using K-means and Gaussian Mixture Model-Expectation Maximization (GMM-EM) algorithms. Both K-means and GMM-EM algorithms can segment lesion area from the rest of brain MRI automatically. The performance metrics (accuracy, sensitivity, specificity, false positive rate, misclassification rate) were estimated for the algorithms and there was no significant difference between K-means and GMM-EM. In addition, lesion size does not affect the accuracy and sensitivity for either method.
Insights
This study compared K-means and Gaussian Mixture Model-Expectation Maximization (GMM-EM) for brain MRI lesion segmentation. Both methods showed similar performance, with lesion size not impacting accuracy or sensitivity.
Area of Science:
- Medical imaging analysis
- Artificial intelligence in healthcare
Background:
- Accurate segmentation of brain lesions in MRI is crucial for diagnosis and treatment monitoring.
- Automated methods are needed to improve efficiency and consistency in lesion segmentation.
Purpose of the Study:
- To compare the performance of K-means and Gaussian Mixture Model-Expectation Maximization (GMM-EM) algorithms for automatic brain MRI lesion segmentation.
- To evaluate the impact of lesion size on the accuracy and sensitivity of these segmentation methods.
Main Methods:
- Brain MRI data were analyzed using K-means clustering and GMM-EM algorithms for lesion segmentation.
- Performance metrics including accuracy, sensitivity, specificity, false positive rate, and misclassification rate were calculated for both algorithms.
Main Results:
- Both K-means and GMM-EM algorithms demonstrated comparable performance in segmenting brain MRI lesions.
- No significant differences in accuracy, sensitivity, or other performance metrics were found between the two methods.
- Lesion size did not significantly affect the accuracy or sensitivity of either K-means or GMM-EM segmentation.
Conclusions:
- K-means and GMM-EM are both effective and comparable automated methods for brain MRI lesion segmentation.
- These algorithms offer reliable performance regardless of lesion size, aiding in consistent clinical assessment.
