Abstract | ||
---|---|---|
•We present an extensive study on topological descriptors in the context of 3D surface classification.•We investigate sensitivity, robustness, and the explanatory power of topological descriptors and compare them with state-of-the-art methods including Convolutional Neural Networks.•Our study shows that topological descriptors can compete with non-topological descriptors and yield a significant improvement of performance when combined with each other.•Our investigation confirms that topological descriptors capture relevant information that is not captured by traditional descriptors. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1016/j.cviu.2017.10.012 | Computer Vision and Image Understanding |
Keywords | DocType | Volume |
Surface texture analysis,3D surface classification,Surface topology analysis,Surface representation,Persistent homology,Persistence diagram,Persistence image | Journal | 167 |
Issue | ISSN | Citations |
1 | 1077-3142 | 4 |
PageRank | References | Authors |
0.43 | 27 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Matthias Zeppelzauer | 1 | 186 | 21.35 |
Bartosz Zielinski 0001 | 2 | 8 | 3.60 |
Mateusz Juda | 3 | 12 | 2.74 |
Markus Seidl | 4 | 30 | 4.36 |