Auto-calibration by locally consistent contours for dental CBCT

Area of Science:

• Medical Imaging
• Radiology
• Computer Vision

Background:

• Dental cone-beam computed tomography (CBCT) image quality is often compromised by inaccurate device calibration.
• Calibration errors and patient motion during scanning can lead to significant artifacts in reconstructed 3D volumes.
• Existing methods may struggle with reconstruction artifacts like truncation.

Purpose of the Study:

• To introduce a novel auto-calibration method for dental CBCT systems.
• To calculate geometrical projection parameters directly from unknown patient geometry.
• To improve the sharpness and reduce artifacts in CBCT images.

Main Methods:

• Developed a novel auto-calibration approach using consistency conditions between projection images.
• Formulated a regularization technique to prevent distortions during optimization.
• Implemented an efficient greedy optimizer for the global optimization problem.

Main Results:

• The auto-calibration method demonstrated robustness against inaccurate initialization.
• The approach is resilient to reconstruction artifacts, including truncation.
• Achieved a relative improvement in sharpness of up to [Formula: see text] in a standard dental CBCT setup.
• Evaluated performance using digital reconstructed radiographs (DRRs) from CT head scans, considering various motion types and anatomical structures.

Conclusions:

• The proposed auto-calibration technique significantly enhances dental CBCT image quality.
• The method offers a robust solution for improving sharpness and reducing artifacts, even in the presence of patient motion.
• This auto-calibration approach provides a valuable tool for optimizing CBCT imaging in dentistry.