摘要
为降低特征识别的复杂度,提出基于特征实体、特征实面和特征虚面概念的层次性特征分类方法.通过构造2类神经网络输入矩阵,利用神经网络在特征识别中所具有的优势,实现基于特征面的分层特征识别方法.实例表明:该方法在识别去除材料的特征时比较有效,但识别特征的范围受到一定限制.
To decrease the complexity of feature recognition,a hierarchy feature classification method based on feature entity,feature concrete face and feature virtual face is proposed.A hierarchy feature recognition method based on feature face is implemented by constructing two kinds of neural network input matrixes,and taking advantage of neural network in feature recognition.The example demonstrates that the method is more effective in recognizing feature of which the material is removed,but the range of feature recognition is somewhat limited.
出处
《计算机辅助工程》
2010年第4期114-117,共4页
Computer Aided Engineering
关键词
特征面
特征识别
神经网络
feature face
feature recognition
neural network
