Characteristics Study on Respiratory Movement of Chest and Abdominal Surface Area for Respiration Tracking in Radiosurgical Robots

Abstract

The traditional methods of representing respiratory movement with external optical markers has insufficient ability of representing the chest and abdominal surface motion, which leads to the low accuracy of tumor motion tracking. To solve this problem, a new respiratory motion representation method based on study of chest and abdominal surface area is proposed. In this paper, point cloud data of chest and abdominal surface during breathing movement is collected by two depth cameras and processed. Processing of the point cloud mainly includes denoising of initial point cloud, registration of multi-frame point cloud, point cloud segmentation and smoothing of the chest and abdominal surface. Then, the Greedy Projection Triangulation algorithm is used to reconstruct the chest and abdominal surface, and the chest and abdominal surface area was calculated. The surface area showed the characteristics of respiratory fluctuation on the whole in the time series. Therefore, the chest and abdominal surface area characteristics can be used to establish the respiratory motion correlation model with the internal tumor for respiration tracking in radiosurgical robots.