Real-time surgical tool tracking and pose estimation using a hybrid cylindrical marker.

Area of Science:

• Medical Imaging
• Robotics in Surgery
• Computer Vision

Background:

• Real-time tracking of surgical tools is crucial for integrated visualization of intraoperative ultrasound and endoscopic images.
• Current tracking methods are limited to planar targets, while most laparoscopic ultrasound probes are cylindrical.
• A robust tracking framework for cylindrical objects is needed to enhance intraoperative ultrasound guidance usability.

Purpose of the Study:

• To develop and validate a novel tracking framework for cylindrical surgical tools.
• To improve the accuracy and reliability of intraoperative ultrasound probe tracking.
• To enhance the integration of ultrasound and endoscopic imaging for surgical guidance.

Main Methods:

• A hybrid marker combining circular dots and chessboard vertices for tracking cylindrical tools.
• Pose estimation using circular dots on the curved surface and chessboard vertices for disambiguation.
• Incorporation of temporal information from consecutive images to minimize tracking failures.
• Real-time performance optimization for computational efficiency.

Main Results:

• The hybrid marker supports a large working space for various tool sizes (6-14 mm diameter).
• The framework achieves high accuracy in translational (40-185 mm) and rotational movements.
• Outperforms existing methods with nearly 100% detection rates and low pose estimation errors (2.8 mm, 0.72°).
• Real-time tracking at 20 frames per second for specified image resolutions.

Conclusions:

• The proposed hybrid marker and tracking framework are effective for a wide range of surgical tools.
• Demonstrates superior detection rates and pose estimation accuracy compared to state-of-the-art methods.
• Applicable for intraoperative ultrasound guidance and general surgical tool tracking in minimally invasive surgery (MIS).