A practical marker-less image registration method for augmented reality oral and maxillofacial surgery

Area of Science:

• Medical Imaging
• Augmented Reality
• Computer-Aided Surgery

Background:

• Image registration is crucial for augmented reality (AR) systems, particularly in surgical navigation where accuracy directly impacts patient outcomes.
• Effective image registration aligns virtual elements with the real surgical field, enhancing visualization and precision.

Purpose of the Study:

• To present a practical, marker-less image registration method for AR-guided oral and maxillofacial surgery.
• To enable accurate overlay of virtual scenes onto real-time surgical video for improved guidance and outcome visualization.

Main Methods:

• Utilized an intraoral 3D scanner to capture patient teeth models intraoperatively.
• Developed a novel 3D stereo matching algorithm for registering the scanner's data with a stereo camera system.
• Employed an iterative closest point scheme to register the 3D model with CT-derived data, linking different spatial references.

Main Results:

• Jaw phantom experiments demonstrated a target registration error below 0.50 mm with a processing time under 0.5 seconds.
• Clinical feasibility was confirmed through volunteer trials, validating the system's practical application.

Conclusions:

• The proposed method is non-invasive, requiring no external fiducial markers.
• Automatic registration ensures scene accuracy, mitigating issues from patient movement.
• This marker-less approach is practical and beneficial for oral and maxillofacial surgery.