摘要
将测试者身体主要部位尺寸作为输入参数,旗袍样板尺寸作为输出参数,逐点训练神经网络,研究其非线性映射关系,并利用神经网络的任意函数逼近功能来获得近似函数关系.采用BP神经网络算法在MATLAB上的实现与应用,建立旗袍样板设计的BP神经网络模型,对旗袍的纸样设计进行研究.通过MATLAB中nntool工具箱对样本进行测试,验证了采用BP神经网络进行旗袍样板设计的有效性和精确性,为MTM中合体服装样板生成系统,提供了理论与技术参考.
Taking the main body parts size of the testers as the input parameters and the pattern sizes of the cheongsam as the output parameters,by training the neural network,studying the nonlinear mapping relationgship of them,and using arbitrary function approximation function of the neural networks,we obtained the approximate function relationship.Established the BP nerve network model of cheongsam pattern design to study the cheongsam pattern design,based on the realization and application of BP（back propagation） neural network algorithm in the MATLAB.The tests for the samples through the application of NNTool toolbox in the MATLAB,verified the feasibility and validity of BP nerve network,and provided a theoretical and technical reference for the automatic generation system of fitting garment pattern in the MTM.
出处
《北京服装学院学报（自然科学版）》
CAS
2013年第1期35-41,共7页
Journal of Beijing Institute of Fashion Technology：Natural Science Edition
