3D Registration of the Point Cloud Data Using Parameter Adaptive Super4PCS Algorithm in Medical Image Analysis

Abstract

In this article, we use the parameter-adaptive Super4PCS algorithm to achieve high-precision registration of medical point clouds. First, generate the corresponding point cloud from the biological data (CT, MRI) to be registered. Then analyze the characteristics of the point cloud to be registered, and use it to adaptively set the parameters of Super4PCS, and finally perform point cloud registration. We compare the performance of six different algorithms with their accuracy and robustness. The accuracy, robustness of our method are the best. At the same time, no parameter input is required which is very convenient for medical workers. Experiments on medical models demonstrate the efficiency and robustness of our algorithm.