Spine Surgery Navigation: Virtual and Real Registration with ARUco Markers

Spine Surgery Navigation: Virtual and Real Registration with ARUco Markers

This study achieved registration between preoperative 3D data of patient spine images and the actual surgical scene by identifying ARUco markers to achieve virtual and real registration. The actual surgical scene was mainly captured by cameras. During the registration process, the relative position between the camera and personalized markers was calculated in real-time. The virtual spine model was placed in the correct position, and the fusion of the virtual spine model and the surgical scene was achieved using computer vision technology. The 3D registration process based on the identification method is shown in Figure 3. Firstly, the video signal obtained by the camera is processed by the computer for identification. Then, the relative position between the camera and the markers is calculated for each frame to obtain the 3D registration matrix. Finally, the virtual spine model is imported into the video stream and displayed in the correct position.

The preoperative registration method is based on ICP registration of personalized markers, which are located in the same position in the patient and preoperative CT images. This method has a fast point selection speed on CT images and patients, and higher accuracy than cloud point registration. The relationship between CT images and personalized markers is shown in Figure 3.

By capturing the position of personalized markers in real space through the camera, we can calculate the transformation matrix. The transformation matrix can be used to convert medical images and models to the position of the actual patient. The transformation matrix is obtained using the least squares (LS) method and singular value decomposition (SVD) algorithm. The objective function of the algorithm is represented as:

	(8)

where R and T represent the rotation matrix and translation matrix, respectively, and they form the transformation matrix. P and Q represent the coordinates of the four corner points of the marker in the R and Q coordinate systems, respectively. The medical images and models are correctly registered to the actual patient through LS and SVD registration. At this point, the intraoperative registration is completed.

In the augmented reality spine surgery navigation system of this paper, the virtual model is dynamically registered in the world space coordinate system by using personalized markers to obtain the transformation matrix, and the spine model is registered and fused with the reconstructed 3D virtual image. The relationship between personalized markers and medical images in the virtual space is obtained through the previous registration work. This method of obtaining the relationship between medical images and patients during the operation avoids the trauma caused by preoperative rigid connections and can better achieve the binding of markers and patients.