Abstract: To recognize interacting feature effectively and improve the recognition efficiency, a
feature recognize algorithm was proposed based on AAG expansion and hint. By simplifying the AAG,
the machining feature faces are considered only for recognition. And then the virtual features are
added to construct the auxiliary constituent surface to complete the separation of interacting features.
The differences of local sub-graph are enhanced by extending the edge properties of the sub-graph,
which improves the recognition efficiency and quality. The feasibility of proposed algorithm was
verified by practical experiment.

Key words: interacting feature, feature recognition, attributed adjacency gragh, sub-graph matching