Spatial anatomic knowledge for 2-D interactive medical image segmentation and matching.

Area of Science:

• Medical Imaging
• Computer Vision
• Computational Anatomy

Background:

• Accurate segmentation of 2D anatomic shapes is crucial for medical image analysis.
• Existing methods often lack efficient ways to incorporate anatomical knowledge, leading to segmentation bottlenecks.
• Representing shape variability within anatomical classes is challenging.

Purpose of the Study:

• To develop a novel representation for 2D anatomic shapes called the radial contour model.
• To implement this model in an interactive image segmentation and matching program (SCANNER).
• To evaluate the model's effectiveness in reducing segmentation time and enabling diagnosis of subtle shape changes.

Main Methods:

• Developed a generic and flexible radial contour model capturing shape class variations.
• Implemented the model in the SCANNER software for 2D image segmentation.
• Trained shape models from 93 contours and evaluated using 60 additional contours against a control model.
• Devised a matching function utilizing the radial contour model.

Main Results:

• The radial contour model reduced interactive segmentation time by a factor of two compared to a control model.
• Significant improvements in segmentation time were observed for specific shapes like the eye.
• The developed matching function demonstrated the potential for diagnosing subtle shape variations.

Conclusions:

• Spatial anatomic knowledge, integrated with interactive tools, can effectively address the segmentation bottleneck in medical imaging.
• The radial contour model provides a flexible and generic representation for anatomical shapes.
• Future extensions to more complex shapes will contribute to a comprehensive anatomical knowledge base for various applications, including image segmentation.