Time-Of-Flight Camera, Optical Tracker and Computed Tomography in Pairwise Data Registration

Area of Science:

• Medical Imaging and Computer-Aided Surgery
• Sensor Fusion and Registration Algorithms

Background:

• Multi-modal data processing is vital for medical applications like minimally invasive surgery.
• Accurate 3D scene analysis and data registration are crucial for computer-aided diagnosis and therapy.
• Time-of-Flight (ToF) sensors offer potential for computer-aided surgery, complementing existing optical tracking systems.

Purpose of the Study:

• To analyze the accuracy of a multi-sensor positioning system integrating pre-operative CT, intra-operative ToF-sensor, and optical tracker data.
• To evaluate the performance of different sensor combinations for 3D scene analysis in a surgical context.

Main Methods:

• Developed a system combining pre-operative CT scans with intra-operative ToF-sensor and optical tracker point clouds.
• Implemented optical sensor setup, ToF-camera calibration, data pre-processing, and Iterative Closest Point (ICP) registration.
• Represented objects as surfaces (CT) or point clouds (ToF, optical tracker).

Main Results:

• Validated registration accuracy using Hausdorff distance and mean absolute distance metrics on human organ phantoms.
• Achieved the best surface alignment with the combination of CT and optical tracker data.
• Observed the lowest accuracy in experiments involving the ToF-camera.

Conclusions:

• The achieved registration accuracies support the continued development of multi-sensor systems for surgical applications.
• Further improvements are needed, particularly concerning the depth information from ToF sensors, for enhanced computer-aided surgery.
• The study provides valuable insights into system limitations and potential advancements for surgical developers.