Abstract | ||
---|---|---|
Geodesic distance has been widely used in building correspondence between two 3D shapes. This paper extends previous work and proposes a correspondence method by geodesic histogram. Geodesic histogram owns two distinct characteristics for correspondence. Firstly, it captures not only the length of two feature points, but also local geometrical features along with the geodesic path. Secondly, it can support partial correspondence since it is locally stable for incomplete 3D shape. Experimental results demonstrate that our algorithm is better able to produce a correct correspondence in the case of pose variation, deformation and even missing parts. Furthermore, statistical analysis shows the proposed method obtains an obvious improvement over the similar method in the literature. |
Year | DOI | Venue |
---|---|---|
2016 | 10.1145/3013971.3013988 | VRCAI |
Keywords | Field | DocType |
3D shape correspondence, Geodesic histogram, Shape diameter function, Partial correspondence | Computer vision,Histogram,Computer science,3d shapes,Histogram matching,Algorithm,Geodesic path,Principal geodesic analysis,Artificial intelligence,Geometry,Geodesic,Statistical analysis | Conference |
Citations | PageRank | References |
0 | 0.34 | 13 |
Authors | ||
4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Xiang Pan | 1 | 8 | 2.51 |
Zhihao Cheng | 2 | 0 | 0.34 |
Junmian Lin | 3 | 0 | 0.68 |
Zhi Liu | 4 | 23 | 14.28 |