Related Experiment Video
Updated: Jul 8, 2026
Automated Segmentation of Cortical Grey Matter from T1-Weighted MRI Images
Published on: January 7, 2019
A modified probabilistic neural network for partial volume segmentation in brain MR image
Tao Song1, Mo M Jamshidi, Roland R Lee
1Man Radiology Department, University of California at San Diego, San Diego, CA 92103, USA. taosong@ucsd.edu
Abstract:
A modified probabilistic neural network (PNN) for brain tissue segmentation with magnetic resonance imaging (MRI) is proposed. In this approach, covariance matrices are used to replace the singular smoothing factor in the PNN's kernel function, and weighting factors are added in the pattern of summation layer. This weighted probabilistic neural network (WPNN) classifier can account for partial volume effects, which exist commonly in MRI, not only in the final result stage, but also in the modeling process. It adopts the self-organizing map (SOM) neural network to overly segment the input MR image, and yield reference vectors necessary for probabilistic density function (pdf) estimation. A supervised "soft" labeling mechanism based on Bayesian rule is developed, so that weighting factors can be generated along with corresponding SOM reference vectors. Tissue classification results from various algorithms are compared, and the effectiveness and robustness of the proposed approach are demonstrated.
Insights
A novel weighted probabilistic neural network (WPNN) improves brain tissue segmentation in MRI by incorporating partial volume effects. This method enhances accuracy and robustness in magnetic resonance imaging analysis.
Area of Science:
- Medical Imaging Analysis
- Artificial Intelligence in Medicine
- Neuroimaging
Background:
- Magnetic resonance imaging (MRI) is crucial for brain tissue segmentation.
- Partial volume effects in MRI introduce complexities in accurate tissue classification.
- Existing probabilistic neural network (PNN) methods have limitations in handling these effects.
Purpose of the Study:
- To propose a modified probabilistic neural network (WPNN) for enhanced brain tissue segmentation in MRI.
- To effectively account for partial volume effects during the segmentation modeling process.
- To improve the accuracy and robustness of MRI-based tissue classification.
Main Methods:
- Developed a weighted probabilistic neural network (WPNN) by replacing the PNN's smoothing factor with covariance matrices.
- Integrated weighting factors into the summation layer of the PNN.
- Employed a self-organizing map (SOM) neural network for image oversegmentation and probabilistic density function (pdf) estimation, coupled with a supervised soft labeling mechanism based on Bayesian rule.
Main Results:
- The proposed WPNN classifier effectively models partial volume effects throughout the segmentation process.
- Demonstrated superior effectiveness and robustness compared to various existing tissue classification algorithms.
- Generated weighting factors alongside SOM reference vectors for improved pdf estimation.
Conclusions:
- The weighted probabilistic neural network (WPNN) offers a significant advancement for brain tissue segmentation in MRI.
- The method's ability to handle partial volume effects enhances the reliability of neuroimaging analysis.
- This approach provides a robust and effective tool for accurate tissue classification in medical imaging.
