Segmentation of Gray Matters and White Matters from Brain MRI data

Abstract

Accurate segmentation of brain tissues such as gray matter and white matter from magnetic resonance imaging is essential for studying brain anatomy, diagnosing neurological disorders, and monitoring disease progression. Traditional methods, such as FSL FAST, produce tissue probability maps but often require task-specific adjustments and face challenges with diverse imaging conditions. Recent foundation models, such as MedSAM, offer a prompt-based approach that leverages large-scale pretraining. In this paper, we propose a modified MedSAM model designed for multi-class brain tissue segmentation. Our preprocessing pipeline includes skull stripping with FSL BET, tissue probability mapping with FSL FAST, and converting these into 2D axial, sagittal, coronal slices with multi-class labels (background, gray matter, and white matter). We extend MedSAM's mask decoder to three classes, freezing the pre-trained image encoder and fine-tuning the prompt encoder and decoder. Experiments on the IXI dataset achieve Dice scores up to 0.8751. This work demonstrates that foundation models like MedSAM can be adapted for multi-class medical image segmentation with minimal architectural modifications. Our findings suggest that such models can be extended to more diverse medical imaging scenarios in future work.

Figures (9)

• Figure 1: Comparison between the original MedSAM model architecture and the modified model architecture.
• Figure 2: Training pipeline comparison: single-orientation models are trained on slices from one anatomical plane (axial, sagittal, or coronal), while the unified model is trained on combined data from all three orientations.
• Figure 3: An example of skull stripping using FSL BET. The left-hand side image shows the original T1-weighted MRI with the skull and non-brain tissues. The right-hand side image shows the result after applying FSL BET. The brain region is cleanly extracted and the surrounding skull is removed.
• Figure 4: An example of tissue segmentation using FSL FAST. The bottom set of images highlights the white matter, while the top set of images highlights the gray matter in the same subject. We have performed the segmentation on skull-stripped images. We have overlaid the tissue boundaries on coronal, sagittal, and axial slices. We have used these maps to construct pixelwise multi-class labels for model training.
• Figure 5: Example slices from the axial plane, coronal plane, and sagittal plane (from left to right)