Abstract: A CAD model retrieval approach based on face context code matching is presented in order to resolve the problem that graph matching or subgraph matching realizing the CAD model retrieval for graph-based model descriptions may not be effective and accurate enough. A representation of Face Adjacency Graph（FAG） for each CAD model is extracted from its B-rep model. According to the face context relation, the FAG is transformed to a layered configuration and face context codes are constructed. The comparison of  face context codes between retrieval model and data models is adopted to exclude and screen the unmatched faces. Experimental results show that this method is able to support CAD model retrieval and its efficiency and accuracy meet the requirement of practical applications.

Key words: face context code, model retrieval, face adjacency graph, subgraph matching

摘要： 为弥补现有的基于图的形状描述应用图匹配或子图匹配实现CAD模型检索时，存在着检索效率低下或检索结果不够准确的缺陷，提出一种基于面上下文码匹配的CAD模型检索方法。提取CAD模型的B-rep信息生成面属性邻接图；依据面的上下文邻接关系，将模型面属性邻接图转化为上下文层状面属性邻接图，生成面上下文码；由检索模型与数据库模型的面上下文码匹配排除筛选不匹配的面，得到与检索模型相匹配的面。实验结果表明，该方法能够实现CAD模型的整体和局部检索，并且检索效率和精准程度能够满足实际需要。

关键词: 面上下文码, 模型检索, 面属性邻接图, 子图匹配