Level set method with automatic selective local statistics for brain tumor segmentation in MR images

Area of Science:

• Medical Imaging
• Image Segmentation
• Computational Biology

Background:

• Image segmentation is crucial for medical diagnosis.
• Magnetic Resonance Imaging (MRI) is widely used for brain imaging.
• Accurate brain tumor segmentation remains a challenge.

Purpose of the Study:

• To propose an automated level set method for brain tumor segmentation in MRI.
• To introduce a novel signed pressure function (SPF) for improved edge detection.
• To enhance the accuracy and robustness of tumor segmentation.

Main Methods:

• Utilized the level set approach for image segmentation.
• Developed a new signed pressure function (SPF) to handle weak or blurred edges.
• Employed local statistics to identify tumor objects within MR images.
• Automated parameter calculation and thresholding for diverse image types.

Main Results:

• The proposed method demonstrated efficient and robust brain tumor segmentation.
• Visual and quantitative evaluations confirmed the algorithm's performance.
• The SPF effectively stopped contours at image edges, improving segmentation.
• Automated parameter adjustment enhanced applicability across different MR images.

Conclusions:

• The developed level set method offers an efficient and robust solution for brain tumor segmentation.
• The novel SPF and automated parameter calculation contribute to improved segmentation accuracy.
• This technique shows promise for clinical applications in neuro-oncology.