A deep learning model for positioning perforated stent. Issue 1 (April 2021)

• Record Type: Journal Article
• Authors: Wang, Erhui; Zheng, Liancun
• Is Part Of: Journal of physics. Volume 1883:Issue 1(2021)
• Publication Date: 2021-04
• Subjects: Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1883/1/012157
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: How to determine the window position of the aortic stent is an important issue in cardiovascular disease research. Currently, this process is viewed in two dimensions, and a lack of information is prone to errors. Here, the real-time frame is suggested to enumerate the three-dimensional shape using the two-dimensional perspective image. A robot-assisted window-opening stent implantation with deep learning is geometrically aligned ratios in window-opening aortic endovascular repair. First, place markers during the windowed stent. Then, we located the three-dimensional pose of each stent segment. Third, the overall three-dimensional shaped stent-graft is achieved by graft gap in terpolation. It has been proposed to separate the markers from the two-dimensional fluorescence microscope for semi-automatic label detection of the image. In this way, we can better understand the opening position of the aortic stent.