Implementation of monocular near-infrared camera tracking in MRI and CT navigation systems

Image-guided intervention (IGI) systems support anatomical registration and real-time instrument tracking during minimally invasive procedures. Existing robotic and navigation solutions are often single-modality, costly, or have limited MRI compatibility. This study presents a cost-effective monocular near-infrared (NIR) tracking system for IGI across multiple modalities while maintaining accuracy and workflow efficiency.

The system, using retroreflective ArUco fiducial markers, was applied to phantom-based percutaneous liver biopsies. Two experiments were conducted under computed tomography (CT) and MRI guidance. Real-time tracking of the needle relative to the phantom and imaging data was performed. An interactive navigation interface was implemented in 3D Slicer. Targeting accuracy and procedure times were evaluated quantitatively.

In the CT-guided study, the target positioning error (TPE) and needle-tip tracking error (TTE) were $10.3\pm 3.7\text{mm}$ and $5.0\pm 1.3\text{mm}$ , respectively, with a total procedure time of $62\pm 34\text{s}$ . In the MRI-guided study, the corresponding values were $4.6\pm 1.0\text{mm}$ , $4.7\pm 2.3\text{mm}$ , and $113\pm 65\text{s}$ , respectively.

The proposed navigation system demonstrates strong potential as an accessible and versatile alternative for image-guided needle interventions across multiple imaging modalities. Furthermore, the system shows promise for future extension to other IGI procedures.