Dense 3D Semantic Mapping of Indoor Scenes from RGB-D Images
Alexander Hermans, Georgios Floros and Bastian Leibe
ICRA 2014 | cited by 146
NYUv1 dataset
Semantic 3D Occupancy Mapping through Efficient High Order CRFs
Shichao Yang∗ Yulan Huang∗ and Sebastian Scherer∗
IROS 2017 | cited by 14 | code | video |
KITTI dataset
Meaningful maps with object-oriented semantic mapping
Niko S¨underhauf1, Trung T. Pham2, Yasir Latif2, Michael Milford1, Ian Reid2
IROS 2017] | cited by 38 |
SceneNet RGB-D dataset
这一篇的意义可能在于直接对三维点进行分割,可以适当参考。但可能这样并不是一个比较好的方式和方法。
DA-RNN Semantic Mapping with Data Associated Recurrent Neural Networks
Yu Xiang and Dieter Fox
IROS 2017] | cited by 24 | code | video | [based on KinectFusion]
dataset: Unsupervised Feature Learning for 3D Scene Labeling
可以参考一下他的depth的设计,实际上是depth aware之前的问题
SemanticFusion: Dense 3D Semantic Mapping with Convolutional Neural Networks
John McCormac, Ankur Handa, Andrew Davison, and Stefan Leutenegger Dyson
ICAR 2017] | cited by 127 | code | video | [Based on ElasticFusion] | [speed about 25Hz]
NYUV2
- Bayesian updates
- 充分利用depth,但是只是简单的把原来网络的三通道变成四通道。第四个通道初始化是rgb的平均值,之后
0–255 colour range to the 0–8m depth range by
increasing the weights by a factor of ≈ 32×.
Semantic Mapping with Simultaneous Object Detection and Localization
Zhen Zeng Yunwen Zhou Odest Chadwicke Jenkins Karthik Desingh
[IROS2018] | [cited by 0] | [code-not open] | [video - no] |
值得参考一下这个, 利用物体检测构建的语义地图
Fast and Accurate Semantic Mapping through Geometric-based Incremental Segmentation
Yoshikatsu Nakajima1, Keisuke Tateno2, Federico Tombari2 and Hideo Saito1
NYUv2
[IROS2018] | [cited by 1] | [code-not open] | [video - no] | [Based on InfiniTAMv3] | [speed about 30Hz]
这一篇当中 II.C 3D geometric segmentation 有很好的总结
Semantic Mapping and Object Detection for Indoor Mobile Robots
S Kowalewsk
[IOP 2019] | [code-not open] | video |
Volumetric Instance-Aware Semantic Mapping and 3D Object Discovery
Margarita Grinvald, Fadri Furrer, Tonci Novkovic, Jen Jen Chung, Cesar Cadena, Roland Siegwart, Juan Nieto
[arXiv:1903.00268] | cited by 1 | [code–not open] | video | [Based on ElasticFusion] | [speed about 1572ms per frame]
sceneNN dataset
PanopticFusion: Online Volumetric Semantic Mapping at the Level of Stuff and Things
Gaku Narita, Takashi Seno, Tomoya Ishikawa, Yohsuke Kaji1
[arXiv:1903.01177] | [code–not open] | video | [Based on ElasticFusion] | [speed about 1572ms per frame]
ScanNet v2
scene0645 01