项目名称: 扩散几何在模型特征分析中应用研究
项目编号: No.61300083
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 自动化技术、计算机技术
项目作者: 王胜法
作者单位: 大连理工大学
项目金额: 23万元
中文摘要: 本项目拟以近年来在数字几何模型处理领域取得成功应用的扩散几何理论为基础,以解决当前模型特征分析中多尺度与多层次特征刻画与检测等关键问题为目的,重点研究高维几何空间中扩散场的构造,模型特征的多尺度及多层次表示,内蕴几何特征检测分析及其相关应用问题。具体的,(1)针对扩散场构造问题,拟从高维几何空间出发,研究具有稀疏性的快速扩散场构造方式,解决传统扩散场计算效率低和对放缩变换鲁棒性差等问题;(2)针对特征表示问题,研究具有多尺度及多层次刻画能力的模型特征描述子,为深层刻画模型特征提供基础工具;(3)针对特征应用问题,研究基于该特征表示的模型内蕴特征检测方案,及其在模型检索、匹配等相关实际问题中的应用。本项目主要涉及扩散几何的基础理论,面向数字几何分析处理领域的核心问题进行应用基础研究,补充和扩展了扩散几何在数字几何模型分析方面的理论,进一步指导其在相关领域中的应用。
中文关键词: 热扩散;几何分析;特征描述子;特征检测;三维打印
英文摘要: Inspired by the successful application of the theory of diffusion geometry in geometric model processing,this project aims at solving the problems in description and detection of multi-scale and hierarchical features. We mainly study the construction of diffusion field in high-dimensional geometry space, the representation of multi-scale and hierarchical features, and the detection of intrinsic features and their related applications. The main contributions include: (1) For the construction of diffusion field, we present an efficient construction of diffusion field with sparsity in a high-dimensional geometry space, which can enhance the computational efficiency and the robustness; (2) For the representation of features, we introduce a multi-scale and hierarchical feature descriptor, which could be a powerful tool for deep depiction of models; (3) For the applications, we rise a scheme to detect intrinsic features, and apply it to model retrieval, model matching and other practical applications. In this project, we make full use of the basic theory of diffusion geometry, devote the efforts on the applied basic research of the core problems of digital geometry processing. This project optimizes and expands the theory of diffusion geometry in digital geometric model analysis, which will further guide its applicati
英文关键词: heat diffusion;geometry analysis;feature descriptor;feature detection;3D Printing