Large-Scale Point Cloud Semantic Segmentation with Superpoint Graphs
By Loic Landrieu, Martin Simonovsky. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2017
1,463 citations recorded by the source corpus.
We propose a novel deep learning-based framework to tackle the challenge of semantic segmentation of large-scale point clouds of millions of points. We argue that the organization of 3D point clouds can be efficiently captured by a structure called superpoint graph (SPG), derived from a partition of the scanned scene into geometrically homogeneous elements. SPGs offer a compact yet rich representation of contextual relationships between object parts, which is then exploited by a graph convolutional network. Our framework sets a new state of the art for segmenting outdoor LiDAR scans (+11.9 and +8.8 mIoU points for both Semantic3D test sets), as well as indoor scans (+12.4 mIoU points for the S3DIS dataset).
Subfields: Computational Mechanics, Environmental Engineering, Geology
View Paper: https://www.semanticscholar.org/paper/7a2e527b6d51071a54aac7a8bdb56ca735a1f78b
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computational Mechanics
- Rohit Girdhar
- Ishan Misra
- Armand Joulin
- Or Litany
- Leonidas J. Guibas
- Chi-Wing Fu
- Li Jiang
- Charles Ruizhongtai Qi
In Environmental Engineering
- Armand Joulin
- Ishan Misra
- Rohit Girdhar
- Li Jiang
- Hengshuang Zhao
- Cyrill Stachniss
- Jens Behley
- Andres Milioto
In Geology
- Or Litany
- Leonidas J. Guibas
- Charles Ruizhongtai Qi
- Jiaya Jia
- Jiwen Lu
- Li Jiang
- Hengshuang Zhao
- Chi-Wing Fu
Citations in the corpus
- PointWeb: Enhancing Local Neighborhood Features for Point Cloud Processing (2019, methodology): “Besides, Superpoint Graph (SPG) [12], on the other hand, focuses on dealing with large point clouds.”
- SqueezeSegV3: Spatially-Adaptive Convolution for Efficient Point-Cloud Segmentation (2020, methodology): “Since a typical outdoor LiDAR (such as Velodyne HDL-64E) can collect about 200K points per scan, it is difficult for previous methods [22,38,47,26,37,31,29] to satisfy a real-time latency constraint.”
- PointSIFT: A SIFT-like Network Module for 3D Point Cloud Semantic Segmentation (2018, background): “Super-point Graphs[12] (SPG) first partitions point cloud into su-perpoints and embed every superpoint with shared Point-Net.”
- ULIP: Learning a Unified Representation of Language, Images, and Point Clouds for 3D Understanding (2022, background): “3D visual understanding research [6, 12, 16, 17, 24, 25] is drawing significant attention in recent years due to the increasing demand of real-world applications such as augmented/virtual reality [1, 27, 31, 46], autonomous driving [23, 57] and robotics [2, 48].”
- An End-to-End Transformer Model for 3D Object Detection (2021, methodology): “Graphs are constructed from 3D data in a variety of ways – DGCNN [77] and PointWeb [90] use local neighborhoods of points, SPG [24] uses attribute and context similarity and Jiang et al.”
- Associatively Segmenting Instances and Semantics in Point Clouds (2019, methodology): “Though great progress has been made in recent years [10, 6, 21, 34, 16] for each single task, no prior method tackles these two tasks associatively.”
- RangeNet ++: Fast and Accurate LiDAR Semantic Segmentation (2019, methodology): “[10] manage to summarize the local relationships in a similar fashion to PointNets by defining a SuperPoint Graph.”
- EC-Net: an Edge-aware Point set Consolidation Network (2018, background): “The network was trained with an edge-aware joint loss function that has four terms; see the yellow boxes on the right. for high-level analysis problems with point clouds [31,32,33,34,35,36,37,38,39,40].”
- SemanticKITTI: A Dataset for Semantic Scene Understanding of LiDAR Sequences (2019, methodology): “To overcome the limitations of the voxel-based representation, such as the exploding memory consumption when the resolution of the voxel grid increases, more recent approaches either upsample voxel-predictions [53] using a CRF or use different representations, like more efficient spatial subdivision”
- RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds (2019, methodology): “Our RandLA-Net takes only 0.04s to directly process a large point cloud with 10 5 points over 150 × 130 × 10 meters in 3D space, which is up to 200 × faster than SPG. Red circles highlight the superior segmentation accuracy of our approach. multilayer perceptrons (MLPs).”
- 4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks (2019, background): “[15] tried to resolve such shortcomings with a recurrent network on top of multiple pointnets, and [16] proposed a variant of 3D continuous convolution for lower layers of a PointNet and got a significant performance boost.”
- PolarNet: An Improved Grid Representation for Online LiDAR Point Clouds Semantic Segmentation (2020, methodology): “The parameterized ones are more commonly seen in the graph-based approaches [34, 32, 15], where the points are modeled as a graph via KNN and then convoluted based on their graph connectivity.”
- Sparse Single Sweep LiDAR Point Cloud Segmentation via Learning Contextual Shape Priors from Scene Completion (2020, methodology): “Among these methods, graph-based learning (Wang et al. 2019a; Landrieu and Simonovsky 2018; Landrieu and Boussaha 2019; Wang et al. 2019c) and convolution-like operations (Thomas et al.”
- Point-GNN: Graph Neural Network for 3D Object Detection in a Point Cloud (2020, background): “[9] partitions a point cloud to simple geometrical shapes and link them into a graph for semantic segmentation.”