Abstract

An effective way to solve the problem of semantic gap is to study of the semantic-based methods in the field of 3D model retrieval. Using multi-channel limited semantic information, we build a heterogeneous semantic network of 3D model, then convert the heterogeneous semantic network to the semantic features. On the base of that, we proposed a method of Semantic-based Laplacian Eigenmap (SBLE). We use the semantically nearest neighbor in heterogeneous semantic network instead of the distantly nearest neighbor in 3D model feature space, and embed the semantic relations in semantic space into the low dimensional feature space by feature mapping. The method retained massive semantic information of 3D models during the course of dimension reduction. Experiment results show the effectiveness of the proposed method for 3D model classification and retrieval on Princeton Shape Benchmark.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call