A novel region-based level set method initialized with mean shift clustering for automated medical image segmentation

Area of Science:

• Medical image analysis
• Computer vision
• Computational imaging

Background:

• Level set methods are crucial for image segmentation but require appropriate initialization and stable evolution.
• Existing methods often struggle with parameter tuning and manual intervention, impacting efficiency and accuracy.

Purpose of the Study:

• To propose a novel region-based level set method that integrates global and local image information for enhanced medical image segmentation.
• To improve the automaticity, efficiency, and accuracy of level set-based segmentation through complementary data utilization.

Main Methods:

• Utilized mean shift clustering to extract global image information and derive appropriate initial contours.
• Employed a data fitting energy to capture local image information for stable evolution of level set curves.
• Integrated global and local information for a complementary approach to region-based segmentation.

Main Results:

• Demonstrated that controlling parameters can be easily estimated from clustering results, simplifying the process.
• Showcased increased automaticity due to reduced computational cost and manual intervention.
• Achieved efficient and accurate medical image segmentation in experimental results.

Conclusions:

• The proposed region-based level set method effectively combines global and local image information.
• The method offers improved automaticity, efficiency, and accuracy in medical image segmentation tasks.
• This approach provides a robust solution for stable contour evolution and initialization in level set methods.