A 3D Shape Automatic Segmentation Method Based on Mean Shift

Abstract

The invention provides a three-dimensional shape automatic partition method based on Mean Shift. Local features of each three-dimensional grid are obtained through distance of a characterization peak, a central coordinate of the grid is obtained through the coordinate of the peak of the three-dimensional-shape grid, and a combined four-dimensional feature space is obtained. Mean Shift algorithm is adopted for conducting cluster calculation in the four-dimensional feature space, and a cluster number and feature points contained in each cluster are obtained. K-nearest neighbor classification technology is adopted for conducting decision space modeling on cluster results and revising partition results partially. Visualization technology is adopted for conducting shading on the partially-revised partition results through a method of marking colors according to cluster attributes. A Princeton partition standard is adopted for calculating standard errors of the partition method under different measurements so as to conduct quantitative evaluation. The three-dimensional shape automatic partition method has the advantages of being high in partition precision and automation and wide in application range of three-dimensional shapes.

Description

technical field [0001] The invention relates to an automatic segmentation method of a three-dimensional shape. Background technique [0002] Mesh segmentation is a key element in the research and application of geometric modeling and computer graphics, assisting in operations such as parameterization, texture mapping, shape matching, deformation, multi-precision modeling, compression, and animation. The understanding of shapes and the acquisition of semantic information based on object representations rely on the extraction of 3D mesh features and structures representing these objects and shapes. Automatic segmentation of 3D surface meshes into functional parts is a basic problem in computer graphics. Segmentation can not only provide semantic information of corresponding objects, but also be used to guide various types of mesh processing algorithms. [0003] In the current domestic and foreign published literature, Marco Attene, Bianca Falcidieno and michela Spagnuolo, "Hi...

AI Technical Summary

Problems solved by technology

[0007] In order to overcome the shortcomings of poor automation performance, small scope of application and low segmentation accuracy in the prior art, the present invention provides a method for automatic segmentation of 3D shapes based on meanshift, which can automatically segment 3D models and CAD models of general objects

Images