Surface Matching-Based Markerless Global Optimization Registration for Improved Optical Surgical Systems in Internet of Things-Enabled Operating Rooms

In the field of smart healthcare, the development of remote surgical treatment technology is of great significance, especially within Internet-of-Things (IoT) enabled operating rooms. In the IoT-enabled operating rooms, the optical surgical navigation system plays a central role in facilitating remote surgical procedures. An essential step in this process is the patient-to-image registration, which serves as a critical requirement for the subsequent image-guided surgery. The prevailing method for patient-to-image registration predominantly relies on artificial landmarks. However, the necessity to attach these landmarks to the patient both during preoperative planning and the entire surgical procedure presents practical inconveniences. As an alternative, markerless registration is employed by aligning two three-dimensional (3D) point sets derived from the patient’s anatomical structure and reconstructed medical images. To acquire a point cloud in the patient space, techniques such as structured light, laser scanning, or Time-of-Flight (TOF) vision are utilized. However, prior research approaches exhibit limitations, including the requirement for reference frames in laser scanning, fixed positioning constraints of structured light systems, and lower precision in point set generation using the TOF vision system. To overcome these limitations, we propose a robust registration method based on surface data from both patient and image domains. Surface points are projected onto the patient’s anatomy through a handheld laser beam, providing a convenient and adaptable projection onto desired positions. This method incorporates a matching algorithm characterized by coarse and fine registration stages, ensuring both global optimization and optimal alignment. We validate the efficacy of our proposed method through comprehensive experiments involving a plastic head phantom and a live human subject. Our experimental results confirm the method’s effectiveness, as demonstrated by an achieved mean target registration error (TRE) of 2.26 mm. This validation highlights the precision and robustness of the method, making it highly suitable for the demands of surgical navigation. Beyond its medical implications, the surgical system introduced by our method holds the potential to contribute to the advancement of smart city infrastructure. In conclusion, the interaction between smart healthcare and IoT technology presents opportunities to revolutionize remote surgical procedures.