摘要
将神经网络方法运用到三维模型的特征识别问题是一种新的尝试 ,对于用神经网络解决拓扑性的、不易被形式化的这类问题具有积极意义。本文综述了近 1 0年来各种基于神经网络的三维模型特征识别技术 ,介绍并分析了三维模型拓扑数据的矢量化方法、不同神经网络模型的识别算法 ,以及基于层分解技术的特征自组织识别等 。
Applying the neural networks(NN) method in 3D model features recognition is challenging, and it is valuable for the future neural network research on other topological or non-formalized problems. This paper supplies a survey of NN methods in 3D model feature recognition in the recent decade. Several preprocessing methods of 3D model data and various NN algorithms are introduced. The most important works include the BP algorithm, AM algorithm and level technology. All of these methods will be helpful to the related researchers.
基金
国家自然科学基金资助项目 ( 6 0 1350 10 )
"973"计划资助项目 (G19980 30 50 9)~~
