Deformable Model Retrieval Based on Topological and Geometric Signatures
IEEE Transactions on Visualization and Computer Graphics
A rectilinearity measurement for 3d meshes
MIR '08 Proceedings of the 1st ACM international conference on Multimedia information retrieval
International Journal of Computer Vision
Parts-based 3D object classification
CVPR'04 Proceedings of the 2004 IEEE computer society conference on Computer vision and pattern recognition
A 3D geometry search engine in support of learning
ICWL'07 Proceedings of the 6th international conference on Advances in web based learning
3D orientation of archaeological fragments coming from a Gothic spire
VAST'04 Proceedings of the 5th International conference on Virtual Reality, Archaeology and Intelligent Cultural Heritage
Hi-index | 0.00 |
Due to the recent improvements in laser scanning technology,3D visualization and modelling, there is an increasingneed for tools supporting the automatic search for 3Dobjects in archives. In this paper we describe a new geometricapproach to 3D shape comparison and retrieval forarbitrary objects described by 3D polyhedral models thatmay contain gaps. In contrast with existing approaches,our approach takes the overall relative spatial location intoaccount by representing the 3D shape as a weighted pointset. To compare two objects geometrically, we generatefor each object a weighted point set, which represents foreach non-empty grid cell a salient point. We compare threemethods to obtain in each grid cell a salient point and aweight: (1) choose the vertex in the cell with the highestGaussian curvature, and choose as weight a measure forthat curvature, (2) choose the area-weighted mean of thevertices in the cell, and choose as weight a measure denotingthe normal variation of the facets in the cell and (3)choose the centre of mass of all vertices in the cell, andchoose as weight one. Finally, we compute the similaritybetween two shapes by comparing their weighted pointsets using a new shape similarity measure based on weighttransportation that is a variation of the Earth Mover's Distance.Unlike the Earth Mover's Distance, the new shapesimilarity measure satisfies the triangle inequality. Thisproperty makes it suitable for use in indexing schemes, thatdepend on the triangle inequality, such as the one we introduce,based on so-called vantage objects. The strengthof our approach is proven through experimental results usinga database consisting of 133 models such as mugs, carsand boats, and a database consisting of 512 models, mostlyair planes, classified into conventional air planes, delta-jets,multi-fuselages, biplanes, helicopters and other models.The results show that the retrieval performance is betterthan related shape matching methods.