Abstract

During TACE interventional surgery, the acquisition and delivery direction of the guidewire's shape is a critical step in the surgery. Traditionally, physicians analyze the shape and endpoint location of the guidewire intraoperatively through real-time X-ray fluoroscopy. However, the low resolution of X-ray images combined with the small proportion that the guidewire occupies within the real-time X-ray image window increases visual fatigue for physicians. When vascular interventional surgical robots undertake the task of guidewire delivery instead of doctors, it is particularly important to analyze the shape and visual feedback of the guidewire. We developed a three-dimensional dynamic model of the guidewire based on MuJoCo, simulating the delivery process of the guidewire in the vascular model during interventional surgery within MuJoCo. The experimental results show that our guidewire model can successfully navigate through different vascular branch structures, completing actions such as advancement, retraction and rotation. This can not only be applied to the training simulation of vascular interventional surgeons but also further explore the vascular interventional surgery robots in conjunction with computer vision technology for auxiliary interventions.

文章链接：https://ieeexplore.ieee.org/document/10633168