Fast Retrieving Approach for 3D Models Based on Semantic Tree

摘要

To solute the problem in 3D model retrieval based on text and content, the application of semantic tree is put forward. The method constructs 3D model semantic tree based on WordNet, under PSB(Princeton Shape Benchmark) ontology library, to proposes a novel approach of constructing semantic tree. It can simplify the process of expansion of semantic tree and adopts amount of information to compute the model which has higher returning similarity. It provides a retrieving method and can optimize the retrieval results according to the feedback mechanism based on shape feature, simultaneously. The experiments show our approach can improve the efficiency of 3D model retrieving effectively.