Segmentation of brain from computed tomography head images

Area of Science:

• Medical Imaging
• Neuroscience
• Computer Vision

Background:

• Accurate segmentation of brain tissues, including gray matter (GM) and white matter (WM), is crucial for neurological studies.
• Computed tomography (CT) head volumes often present challenges for segmentation due to large slice thickness and image noise.
• Existing methods may struggle with precise delineation of brain structures in low-resolution CT data.

Purpose of the Study:

• To develop and validate an algorithm for segmenting human brain matter (GM and WM) from CT head volumes with large slice thickness.
• To improve the accuracy of brain segmentation in CT imaging, particularly in cases with pathological findings.
• To provide a robust method for automated brain tissue identification in clinical and research settings.

Main Methods:

• A novel algorithm combining thresholding and brain mask propagation for CT brain segmentation.
• Utilizes a 2D reference image for intensity characteristics and Fuzzy C-means clustering for threshold determination.
• Employs a distance criterion for identifying brain candidates and mask propagation for final brain identification.

Main Results:

• The algorithm successfully determines thresholds for head mask and brain segmentation.
• Brain candidates are identified using a distance criterion, followed by mask propagation for accurate brain identification.
• Validation against non-enhanced and enhanced CT volumes with pathology demonstrated the algorithm's effectiveness.

Conclusions:

• The proposed algorithm effectively segments human brain matter (GM and WM) from CT volumes with large slice thickness.
• The method shows promise for accurate brain segmentation in the presence of pathology.
• This technique offers a valuable tool for analyzing brain structures in CT imaging.