Abstract | ||
---|---|---|
The rapid development of 3D acquisition devices enables us to collect billions of points in a few hours. However, the analysis of the output data is a challenging task, especially in the field of 3D point cloud change detection. In this Shape Retrieval Challenge (SHREC) track, we provide a street-scene dataset for 3D point cloud change detection. The dataset consists of 866 3D object pairs in year 2016 and 2020 from 78 large-scale street scene 3D point clouds. Our goal is to detect the changes from multi-temporal point clouds in a complex street environment. We compare three methods on this benchmark, with one handcrafted (PoChaDeHH) and the other two learning-based (HGI-CD and SiamGCN). The results show that the handcrafted algorithm has bal-anced performance over all classes, while learning-based methods achieve overwhelming performance but suffer from the class-imbalanced problem and may fail on minority classes. The randomized over -sampling metric applied in SiamGCN can alleviate this problem. Also, different siamese network archi-tecture in HGI-CD and SiamGCN contribute to the designing of a network for the 3D change detection task. (c) 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ) |
Year | DOI | Venue |
---|---|---|
2021 | 10.1016/j.cag.2021.07.004 | COMPUTERS & GRAPHICS-UK |
Keywords | DocType | Volume |
SHREC 2021, 3D Point cloud change detection, Graph convolutional networks, Siamese networks | Journal | 99 |
ISSN | Citations | PageRank |
0097-8493 | 0 | 0.34 |
References | Authors | |
0 | 9 |
Name | Order | Citations | PageRank |
---|---|---|---|
Tao Ku | 1 | 0 | 0.34 |
Sam Galanakis | 2 | 0 | 0.34 |
Bas Boom | 3 | 0 | 0.34 |
Remco C. Veltkamp | 4 | 2127 | 157.19 |
Darshan Bangera | 5 | 0 | 0.34 |
Shankar Gangisetty | 6 | 2 | 2.43 |
Nikolaos Stagakis | 7 | 0 | 0.34 |
Gerasimos Arvanitis | 8 | 9 | 6.21 |
K. Moustakas | 9 | 285 | 58.02 |