A multiresolution binary level set method and its application to intracranial hematoma segmentation.

Area of Science:

• Medical image analysis
• Computational imaging
• Biomedical engineering

Background:

• Image segmentation is crucial for medical diagnosis.
• Existing binary level set methods struggle with irregular image margins and efficiency.
• Accurate segmentation of intracranial hematomas is clinically significant.

Purpose of the Study:

• To develop a robust multiresolution binary level set method for image segmentation.
• To enhance the edge length approximation for irregular image margins.
• To improve segmentation efficiency and accuracy for intracranial hematomas.

Main Methods:

• A modified binary level set formulation based on the Song-Chan algorithm was employed.
• Edge length approximation was adapted to handle irregular image margins.
• Image pyramids were utilized for multiresolution processing, enhancing computational efficiency.

Main Results:

• The proposed method successfully segments objects near image margins.
• It enables efficient segmentation of intracranial hematomas on brain CT slices.
• Segmentation of epidural and subdural hematomas was achieved with accuracy comparable to human experts.

Conclusions:

• The multiresolution binary level set method offers an efficient and accurate solution for medical image segmentation.
• This technique successfully addresses limitations of previous methods, particularly for irregularly shaped structures.
• It demonstrates significant potential for clinical applications in neuroimaging.