3d Similarity Search Using A Weighted Structural Histogram Representation

Tong Lu,Rongjun Gao,Tuantuan Wang,Yubin Yang
DOI: https://doi.org/10.1007/978-3-642-15702-8_32
2010-01-01
Abstract:A fast and robust 3D retrieval method is proposed based on a novel weighted structural histogram representation. Our method has the following steps: 1) adaptively segment any 3D shape into a group of meaningful parts to generate local distribution matrixes, 2) integrate all the local distribution matrixes into a global distribution matrix, simultaneously considering their weight factors, and 3) retrieve 3D shapes by calculating the distance between their global distribution matrixes. Experimental results show that our method is effective and efficient for 3D shape retrieval and robust to translation, scaling or rotation transformations.
What problem does this paper attempt to address?