architecture

Research on the method of anti-occlusion of surgical instrument tracking based on multi-camera module information fusion



https://doi.org/10.1016/j.measurement.2024.115480Get rights and content

Highlights

  • A surgical instrument tracking block composed of multi-object is designed.
  • A method for calculating the end position of surgical instruments is proposed.
  • This paper presents an improved method of motion blur for surgical instruments.
  • An algorithm of information fusion to track surgical instruments are proposed.

Abstract

The traditional optical tracking system is easily affected by occlusion and a narrow field of vision, which leads to the failure of surgical instrument tracking. To solve this problem, a multi-target tracking method based on multi-camera module information fusion is proposed. Firstly, a multi-target tracking block is installed at the end of the surgical instrument. A method of solving the end position of surgical instruments based on multi-object tracking is proposed to improve the success rate of surgical instrument tracking. Secondly, motion blur is easy to occur in surgical instrument tracking. A method to reduce the tracking failure caused by motion blur is proposed to improve the sharp edge of the image. Then, a method of information fusion of multi-camera module is proposed to track surgical instruments. Finally, experiments are designed to verify the stability and accuracy of the proposed method. The tracking accuracy is 1.89 ± 0.24 mm.

Introduction

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.

Access through your organization

Check access to the full text by signing in through your organization.

Section snippets

Related work

Many studies have begun to introduce multi-camera system for tracking, and multi-camera system are gradually being promoted and applied in clinical surgery. Vörös et al. [19] developed a multi-camera tracking system based on clinical operation workflow. The system consists of multiple cameras and Aruco markers, which are used to track the surgeon and monitor the surgery. In order to solve the problem of surgeons' inability to monitor 3D anatomical areas in a narrow range, Faragasso et al. [20]

Methodology

  • A.
    Proposed system
The traditional optical tracking system has the problem of occlusion and narrow field of view. We have built a surgical instrument tracking system based on multi-camera module information fusion, as shown in Fig. 2(a) and (c). Considering the current imbalance of resources in hospitals, our proposed system is mainly targeted at hospitals with scarce medical resources. The greatest advantage of our proposed system is its relatively low cost while meeting the basic needs of

Experiments and analysis of results

This paper uses four monocular tracking modules to form a multi-module tracking system for surgical instruments. As a surgical instrument tracking system, the most concerned indexes of surgeons are the accuracy and stability of the system. Therefore, experiments are designed to verify the tracking accuracy and stability of the system according to these indexes.
The illumination intensity is often an essential factor affecting the accuracy of an optical tracking system. The tracking accuracy

Discussion

This paper proposes a method to solve the occlusion problem of optical tracking systems by combining multi-object recognition with a multi-camera tracking module. We propose two different methods for multi-camera module information fusion to track and locate the position of surgical instruments during surgery. In addition, we compare them with the existing EDPM and RFAM. The tracking accuracy of our proposed AWPM method can reach 1.89 ± 0.24 mm.
The above experiments verify the tracking accuracy 

Conclusion

This paper proposes multi-module data fusion based on the previous method of multi-camera tracking. This method does not require complex and expensive hardware facilities. At the same time, it can inhibit the motion blur and improve the tracking accuracy of surgical instruments. In addition, a method is proposed to reduce motion blur by reducing noise and improving the sharp edge of the image. In this way, the failure rate of surgical instrument tracking due to motion blur can be reduced. The