Point-GNN: Graph Neural Network for 3D Object Detection in a Point Cloud

By Weijing Shi, R. Rajkumar. Computer Vision and Pattern Recognition, 2020

921 citations recorded by the source corpus.

In this paper, we propose a graph neural network to detect objects from a LiDAR point cloud. Towards this end, we encode the point cloud efficiently in a fixed radius near-neighbors graph. We design a graph neural network, named Point-GNN, to predict the category and shape of the object that each vertex in the graph belongs to. In Point-GNN, we propose an auto-registration mechanism to reduce translation variance, and also design a box merging and scoring operation to combine detections from multiple vertices accurately. Our experiments on the KITTI benchmark show the proposed approach achieves leading accuracy using the point cloud alone and can even surpass fusion-based algorithms. Our results demonstrate the potential of using the graph neural network as a new approach for 3D object detection. The code is available at https://github.com/WeijingShi/Point-GNN.

Subfields: Computational Mechanics, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/22cc76c6d9b25facb2874bbcbbbfe781a4d85bcd

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

1. Yulan Guo
2. Li Jiang
3. Chi-Wing Fu
4. Gage DeZoort
5. Can Qin
6. Weiliang Tang
7. Shiguo Lian
8. Gao Huang

In Computer Vision and Pattern Recognition

1. Yun Raymond Fu
2. Mohammed Bennamoun
3. Shiji Song
4. Yulan Guo
5. Gao Huang
6. Xuran Pan
7. Chi-Wing Fu
8. Li Liu

Citations in the corpus

• Instance Segmentation GNNs for One-Shot Conformal Tracking at the LHC (2021, methodology): “Our conformal GNN architecture is inspired by the Point-GNN architecture for bounding box-based instance segmentation [2].”
• Vision-based robotic grasping from object localization, object pose estimation to grasp estimation for parallel grippers: a review (2020, methodology): “Point-GNN [Shi and Rajkumar, 2020] utilized graph neural network on the point cloud and designed a graph neural network with an auto-registration mechanism which detects multiple objects in a single shot.”
• Deep Learning for 3D Point Clouds: A Survey (2019, background): “[125] proposed a graph neural network Point-GNN to detect 3D objects from lidar point clouds.”
• ARKitScenes: A Diverse Real-World Dataset For 3D Indoor Scene Understanding Using Mobile RGB-D Data (2021, background): “3D object detection is a computer vision task that has gained a lot of popularity in recent years [27, 28, 29, 30, 31, 32, 33, 4, 5, 34, 7].”
• 3D Object Detection for Autonomous Driving: A Survey (2021, methodology): “As shown in Table 11, Point-GNN [52], SECOND [25], 3DSSD [51] Part-A 2 [56], SA-SSD [53] and CIA-SSD [26] maintain a reasonable accuracy when LiDAR signals are reduced from 64 to 32.”
• Learning Distilled Collaboration Graph for Multi-Agent Perception (2021, background): “Comprehensive experiments conducted in 3D object detection [37, 28, 29, 17] have shown that the proposed DiscoNet achieves better performance-bandwidth trade-off and lower communication latency than the state-of-the-art intermediate collaboration methods.”
• CIA-SSD: Confident IoU-Aware Single-Stage Object Detector From Point Cloud (2020, background): “Also, the detection precisions of recent single-stage detectors (He et al. 2020; Yang et al. 2020; Shi and Rajkumar 2020) gradually approach that of the state-of-the-art two-stage detectors.”
• Graph Neural Networks for Intelligent Transportation Systems: A Survey (2023, background): “Jin et al. [108] argue that applying convolution operators to point clouds is not efficient because point clouds are not evenly distributed over grids.”
• PV-RCNN++: Point-Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detection (2021, neutral): “14 Point-GNN [92] CVPR 2020 LiDAR only 88.”
• Rethinking Network Design and Local Geometry in Point Cloud: A Simple Residual MLP Framework (2022, background): “…in various applications, including 3D shape classification (Qi et al., 2017a), semantic segmentation (Hu et al., 2020) and object detection (Shi & Rajkumar, 2020), etc. Recent efforts have shown promising results for point cloud analysis by exploring local geometric information, using…”
• Not All Points Are Equal: Learning Highly Efficient Point-based Detectors for 3D LiDAR Point Clouds (2022, neutral): “PointGNN [40] is a framework by generalizing graph neural network to 3D object detection.”
• SE-SSD: Self-Ensembling Single-Stage Object Detector From Point Cloud (2021, methodology): “Point-GNN [26] exploits a graph neural network to learn point features.”
• 3D Object Detection with Pointformer (2020, methodology): “Detectors are designed either with point clouds as the only input [42, 41, 15, 39, 24, 25, 19, 26, 35, 40] or fusing multiple sensor modalities such as LiDAR and camera [20, 17, 31].”