基于CHNN聚类算法的款式部件生成模型

对智能化服装款式设计系统中的款式部件的自动获取功能进行了研究.采用基于连续Hopfield神经网络(CHNN)的聚类算法提出了一个款式部件的风格生成模型.提取表现部件造型特征的特征要素构造一个空间点集,利用CHNN网络对该点集进行聚类,分析部件类别与款式设计风格之间的关系,建立基于款式风格设计的部件搭配规则.并将该模型应用于款式的衣片部件上,实现了衣片部件的聚类.实验结果表明,该模型设计合理,分类清晰,具有可扩展性.

Part style developing model based on CHNN clustering algorithm

The auto-gained function of parts in the intelligent fashion design system is studied. A method based on continuous Hopfield neural network (CHNN) clustering algorithm is used to accomplish a part style developing model. Some characteristic variables which represented the property of part structure are gained to make up of a space-dot set. Then the set is classified by CHNN. By analyzing the relations between the part sorts and garment styles, the rules of part arrangement are gained. The model is used on the part named coat piece and the classification is realized. According to the experimental results, the model is expansive and the method is effective.