纺织学报 ›› 2008, Vol. 29 ›› Issue (1): 129-132.

• 管理与信息化 • 上一篇    

基于神经网络的服装制板系统

丛芳;赵野军   

  1. 北京服装学院计算机信息中心 北京100029
  • 收稿日期:2006-07-19 修回日期:2006-09-25 出版日期:2008-01-15 发布日期:2008-01-15

Realization of the intelligent garment pattern system based on Neural Networks

CONG Fang;ZHAO Yejun   

  1. Conputer Information Center Beijing Institute of Clothing Technology;Beijing 100029;China
  • Received:2006-07-19 Revised:2006-09-25 Online:2008-01-15 Published:2008-01-15

摘要: 利用智能技术来打破服装参数化制板的传统模式,以女装原型数据为对象,Java作为编程语言,对BP神经网络进行训练并检验。仿真结果表明:该算法在制作服装板型上完全可以达到与人脑相似的程度;在板型数据提取时,改变以往提取长度数据的计算模式,通过设定参数坐标,使图形与数据之间的相互转换更为方便;同时参数坐标数值在BP网络的训练中也表现了优异的包容性和数据相关性。

Abstract: This paper maked a breakthrough of the traditional mode of the garment parameterization pattern method.Taking the prototype women′s garment as the study object,the BP neural network was trained and verified with Java.The simulation results indicated that this algorithm was similar with our brain on making garment pattern.And this paper used the method of parametric coordinate which differs from those of extracting the length of pattern in the past.In this way,it was easier to extract the data from the pattern and to make transfer between them.These coordinate data express good data-correlation in the training of BP neural networks.

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!