Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (21): 174-178.DOI: 10.3778/j.issn.1002-8331.2008.21.048

• 机器学习 • Previous Articles     Next Articles

Analyzing 3D model visual features based on neural networks

WEI Wei1,2,YANG Yu-bin1,LIN Jin-jie1,RUAN Jia-bin1   

  1. 1.State Key Laboratory for Novel Software Technology,Nanjing University,Nanjing 210093,China
    2.Software Institute,Nanjing University,Nanjing 210093,China
  • Received:2008-04-30 Revised:2008-05-28 Online:2008-07-21 Published:2008-07-21
  • Contact: WEI Wei

基于神经网络的三维模型视觉特征分析

韦 伟1,2,杨育彬1,林金杰1,阮佳彬1   

  1. 1.南京大学 计算机软件新技术国家重点实验室计算机科学与技术系,南京 210093
    2.南京大学 软件学院,南京 210093
  • 通讯作者: 韦 伟

Abstract: A review on feature descriptors,including shape,color,texture and material,is firstly given in this paper.Then a three-dimensional visual vector is designed for neural works to classify 3d models.Classification methods based on perceptron neural network and Hopfield neural network are proposed.Experimental results have shown that,the proposed methods are able to simulate human’s visual perception effectively and efficiently.Finally,conclusions are drawn and the future work on fuzzy neural network is introduced.

Key words: 3D model, visual feature, perceptron, Hopfield, Neural Network classification

摘要: 首先从形状、颜色、纹理材质三个主要视觉特性入手,阐述模型的特征描述符,设计三元组视觉特征向量用于神经网络进行模型分类。具体基于感知器神经网络、Hopfield神经网络分别实现了对三维物体的分类。实验表明,基于神经网络的分类器能对基于视觉特征描述的三维物体进行有效识别。

关键词: 三维模型, 视觉特征, 感知器神经网络, Hopfield网络, 三维物体分类