Multiresolutional similarity assessment and retrieval of solid models based on DBMS

Similarity assessment and retrieval of solid models play a key role in achieving reuse of huge amounts of solid models. In this paper, an approach to multiresolutional similarity assessment and retrieval of solid models is presented. First the dilation based multiresolutional skeleton of a solid model is defined, which is adopted as the shape descriptor of solid models in the approach; Then the detailed algorithm of generating the proper DBMS of a solid model through multiresolutional dilation is described; After that the methods of multiresolutional matching based on the DBMS graph and the multiresolutional similarity assessment and retrieval of solid models based on multiresolutional matching are presented; Finally some preliminary test results are given.