Sample-Based Complex 3D Data Generation Technology

In this paper,we present a sample-based complex three-dimensional data generation technology,and the whole process is divided into three main stages:data acquisition and preprocessing,analysis and modeling,data generation.The use of sample-based generation technology,on the one hand can get rid of the tedious parameters setting,on the other hand can ensure that the generated results contain samples' important features.On this basis,we realize infinite terrain generation based on real-world elevation data and vessels generation based on MRI data.In the terrain generation technology,the value noise is used to generate the terrain height.The value noise can be easily combined with the height distribution of the sample,so that the generated terrain has a similar height distribution with the sample.In order to improve the rendering performance and effectiveness of the generated terrain,we also use GPU tessellation and LOD to generate a terrain with infinite range and detail.In vessels generation technology,we propose a method of sample parameterizing.Through statistical analysis of the samples,we construct a three-level tree-like statistical model,and then the blood vessels are generated by using the statistical model.Through user surveys,we verify the realism of the generated vessels.