Research On Liver Surgery Navigation System Based On Augmented Reality

During the traditional liver tumor resection surgery,doctors often compar preperative two-dimensional images generated by CT or MRI and patient’s organs to locate important anatomical tissues and structures of liver such as tumor and blood vessels.This traditional two-dimensional image-guided operation has a high risk of surgery and requirement for doctor’s clinical experience.Besides,it is diffcult to understand two-dimensional information such as CT,MRI and ultrasound images.The above problem limit the development of liver operations.However,with the development of computer graphics and image technogy,this change has a new solution.Through augmented reality technology,the three-dimensional model reconstructed before the operation is projected into the operation scene,which can intuitively display the internal structure of the patient’s liver and guide the doctor to perform the operationBased on the analysis and explanation of the current status of liver surgery navigation system at home and abroad,this article conducts in-depth research and provides solutions for two types of problems of existing navigation systems,namely augmented reality schemes and deformation registration problems.The core of augmented reality is to unify the virtual model and the real model under the same coordinate system.This article simulates the camera’s imaging process by studying the camera’s pinhole imaging model and virtual imaging technology,and builds a virtual camera with the same imaging results as the real camera.The virtual image generated by the virtual camera is superimposed with the surgical scene shot by the real camera to obtain an augmented reality picture with instructive significance.In order to obtain the final augmented reality picture,generating the correct virtual picture is particularly critical.And spatial locating technology is the core of it.In this article,the camera calibration technology and infrared-based locating and tracking system are used to determine the spatial pose of the liver relative to the camera,and the relative pose relationship in the real space is backcalculated into the virtual space.At the same time,the internal parameters of the real camera are assigned to the virtual camera.The projection matrix is finally generated as a virtual image.In order to obtain the position of the liver relative to the camera in real time during the operation,the liver point cloud model reconstructed by CT before the operation is aligned with the liver point cloud surface obtained by the three-dimensional structured light scanner,and the position of the liver model in the camera coordinate system is solved.posture.In this paper,the iterative nearest point algorithm is used as the fine registration algorithm for the point cloud.Since the iterative nearest point algorithm requires a good initial pose,the random consistent sampling described by the fast point feature histogram is used as the coarse registration,and finally the liver is achieved.Real-time non-contact locating.Since the liver is a soft tissue,the difference in the posture of the patient during the preoperative CT scan and the actual operation will lead to the difference in the shape of the liver model during the preoperative operation,which makes the final registration error too large.In order to compensate for the error caused by liver deformation,position dynamics is introduced,and a non-rigid registration algorithm based on physical deformation model is proposed.Finally,in order to improve the calculation efficiency of the deformed model,and the parallel calculation program is optimized by the coloring algorithm,which improves the calculation efficiency of the program.In view of the above content,the liver surgery navigation is built,and the experimental process is introduced in detail.Finally,the source and composition of the error are analyzed.The results show that the system has a higher guiding significance for liver clinical surgery.