Vertebrae localization in CT using both local and global symmetry features.

Area of Science:

• Medical Imaging
• Computer-Aided Diagnosis
• Surgical Planning

Background:

• Accurate vertebrae segmentation and localization in CT images are critical for medical applications.
• Vertebrae present significant localization challenges due to their complex shapes, deformations, and occlusions.

Purpose of the Study:

• To develop an improved method for precise vertebrae localization in CT images.
• To leverage both local appearance and symmetry features for enhanced localization accuracy.

Main Methods:

• Incorporation of local appearance features with global translational and local reflection symmetry features.
• Proposal of a Sphere Surface Expansion method for efficient 3D reflection symmetry investigation.
• Development of an iterative optimization framework for localization.

Main Results:

• The proposed method demonstrates superior performance compared to existing localization techniques.
• Quantitative and qualitative evaluations confirm the effectiveness of the symmetry-based approach.

Conclusions:

• Integrating symmetry features significantly improves vertebrae localization accuracy in CT images.
• The proposed method offers a robust solution for challenging vertebrae localization tasks in medical imaging.