WITH the continuous improvement of the medical level, the surgical navigation system has become an indispensable part of liver surgery [1], [2]. Computer-aided surgical navigation can improve surgeons' visual information and surgical procedures [3]. However, in the course of liver surgery, dynamic changes can occur in the anatomic region of liver operation. It means surgical planning must be done promptly through changes in surgical instruments. The tracking of surgical instruments can reveal the relative position relationship between the surgical instruments and the anatomical structure of the patient's liver. Therefore, it is important to track the location of surgical instruments in real time [4], [5].
The surgical tracking system that can be realized commercially mainly includes an electromagnetic tracking system, a mechanical positioning system, and an optical tracking system [6]. An electromagnetic tracking system mainly generates a magnetic field through the magnetic transmitter. The position of the sensors mounted on the surgical instruments can then be detected to enable tracking of the surgical instruments. This method can solve the common problem of visual occlusion in the optical tracking system [7].
However, this method is highly susceptible to electromagnetic interference, and the scope of the surgical area tracked is relatively limited. Its complex mechanical structure often limits tracking and is challenging during actual operation. The optical tracking system is the most widely used surgery, mainly using light to triangulate objects [8]. Generally speaking, it has high accuracy. Optical tracking system mainly includes infrared or laser and video measurement system. The infrared tracking system requires the installation of infrared reflective spheres on the surgical instruments, which are tracked by infrared cameras. [9]. This method has high tracking accuracy, but the cost is relatively high [10]. A tracking system based on ordinary optical cameras requires the design of markers that can be attached to surgical instruments, and the camera can track surgical instruments.
NDI optical tracking system is a widely used surgical tracking system [11]. The Micron Tracker optical tracking system based on binocular vision is also used in surgery [12]. In addition, it also includes some standard optical tracking systems such as Krios [13], Optotrak Certus [14], etc. As shown in Fig. 1, optical tracking systems currently available on the market are demonstrated. These optical tracking systems are generally equipped with more expensive optical components. These components can improve the tracking accuracy of optical systems.
However, although these optical tracking systems have been widely used in surgical procedures, there are still some disadvantages. Currently, most optical tracking systems adopt the principle of binocular stereo vision. The most significant advantage of these systems is that the tracking algorithm is simple [15], [16]. However, surgical instrument tracking is not possible when one of the cameras is blocked. In the operating room, optical tracking systems are often manipulated by the surgeon's position, resulting in failure to track surgical instruments [17], [18]. In addition, the optical tracking system based on fiducial markers often produces motion blur due to the movement of surgical instruments. This will reduce the accuracy of surgical instrument tracking and increase the risk of surgery.
The traditional optical tracking system has the problems of poor occlusion resistance and narrow field of vision. This paper proposes a method based on multi-camera module information fusion tracking to solve the problem. In addition, the paper provides a solution to motion blur in tracking surgical instruments. The feasibility and robustness of the proposed method are verified by experiments. The main contributions of this paper include the following:
- 1.
A tracking block composed of multiple fiducial markers is fixed at the end of surgical instruments to improve the success rate of monocular module tracking. The method of solving the end position of surgical instruments based on multiple fiducial markers is proposed.
- 2.
In order to reduce the tracking failure caused by motion blur, a method is proposed to improve the sharp edge of the image by reducing the noise.
- 3.
The multi-camera information fusion tracking module is composed of several monocular tracking modules. In addition, different multi-module information fusion algorithms are compared to solve the problem of occlusion during surgical instrument tracking.
- 4.
Experiments on surgical instrument tracking were designed to verify the robustness and effectiveness of the proposed method.
This paper is organized as follows: Section 2 briefly reviews and summarizes the latest progress in optical tracking systems from multi-camera module tracking, motion blur, etc. Section 3 introduces the system overview of multi-camera information fusion. It also proposes the methods of solving the motion blur of surgical instruments and the end position of surgical instruments based on multi-object, respectively. Section 4 experiments are designed to verify the effectiveness and feasibility of the proposed method. The experimental results are analyzed and compared with the existing methods. Section 5 discusses the differences between the proposed method, the current mainstream optical tracking systems, and other influential factors. Section VI concludes and points out directions for further research.