[Paper Review] Brain MRI Segmentation using Rule-Based Hybrid Approach

Mustansar Fiaz, Kamran Ali+3 more|arXiv (Cornell University)|Feb 12, 2019

Brain Tumor Detection and Classification19 references4 citations

TL;DR

This paper proposes a rule-based hybrid approach for brain MRI segmentation that combines Gabor feature extraction with multiple classifiers—ISNN, KNN, PNN, and SVM—selecting the optimal classifier per tissue type based on performance. The method achieves improved segmentation accuracy across different brain tissues by dynamically assigning the best-performing classifier per tissue, outperforming individual classifiers in experimental evaluations.

ABSTRACT

Medical image segmentation being a substantial component of image processing plays a significant role to analyze gross anatomy, to locate an infirmity and to plan the surgical procedures. Segmentation of brain Magnetic Resonance Imaging (MRI) is of considerable importance for the accurate diagnosis. However, precise and accurate segmentation of brain MRI is a challenging task. Here, we present an efficient framework for segmentation of brain MR images. For this purpose, Gabor transform method is used to compute features of brain MRI. Then, these features are classified by using four different classifiers i.e., Incremental Supervised Neural Network (ISNN), K-Nearest Neighbor (KNN), Probabilistic Neural Network (PNN), and Support Vector Machine (SVM). Performance of these classifiers is investigated over different images of brain MRI and the variation in the performance of these classifiers is observed for different brain tissues. Thus, we proposed a rule-based hybrid approach to segment brain MRI. Experimental results show that the performance of these classifiers varies over each tissue MRI and the proposed rule-based hybrid approach exhibits better segmentation of brain MRI tissues.

Motivation & Objective

• To address the challenge of accurate and robust segmentation of brain MRI tissues, which is critical for diagnosis and surgical planning.
• To evaluate the performance of four classifiers—ISNN, KNN, PNN, and SVM—on brain MRI segmentation tasks.
• To develop a hybrid approach that selects the optimal classifier per tissue type based on empirical performance to enhance overall segmentation accuracy.
• To reduce variability in segmentation results across different brain tissues by leveraging classifier-specific strengths.

Proposed method

• Gabor transform is applied to extract texture and spatial frequency features from brain MRI scans.
• Four classifiers—Incremental Supervised Neural Network (ISNN), K-Nearest Neighbor (KNN), Probabilistic Neural Network (PNN), and Support Vector Machine (SVM)—are trained and evaluated on the extracted features.
• Performance of each classifier is assessed separately on different brain tissues (e.g., white matter, gray matter, CSF) to identify tissue-specific optimal performance.
• A rule-based system is designed to dynamically select the best-performing classifier for each tissue type based on experimental results.
• The hybrid framework integrates the selected classifiers to produce a final, unified segmentation output.
• The approach is validated on a dataset of brain MRI images, with performance measured using standard segmentation metrics.

Experimental results

Research questions

• RQ1 Which classifier among ISNN, KNN, PNN, and SVM performs best for each type of brain tissue in MRI segmentation?
• RQ2 How does classifier performance vary across different brain tissues (e.g., white matter, gray matter, cerebrospinal fluid)?
• RQ3 Can a rule-based fusion strategy combining multiple classifiers improve overall segmentation accuracy compared to individual classifiers?
• RQ4 Does dynamic classifier selection based on tissue-specific performance lead to more consistent and accurate segmentation results?

Key findings

• The performance of individual classifiers varies significantly across different brain tissues, with no single classifier outperforming all others across all tissue types.
• The proposed rule-based hybrid approach achieves superior segmentation accuracy compared to any single classifier by selecting the best-performing classifier per tissue.
• Gabor features effectively capture texture and structural information useful for distinguishing brain tissue types in MRI.
• The hybrid system reduces segmentation error by leveraging the strengths of multiple classifiers tailored to specific tissue characteristics.