Voting for Voting in Online Point Cloud Object Detection

By Dominic Zeng Wang, Ingmar Posner. 2015

378 citations recorded by the source corpus.

This paper proposes an efficient and effective scheme to applying the sliding window approach popular in computer vision to 3D data. Specifically, the sparse nature of the problem is exploited via a voting scheme to enable a search through all putative object locations at any orientation. We prove that this voting scheme is mathematically equivalent to a convolution on a sparse feature grid and thus enables the processing, in full 3D, of any point cloud irrespective of the number of vantage points required to construct it. As such it is versatile enough to operate on data from popular 3D laser scanners such as a Velodyne as well as on 3D data obtained from increasingly popular push-broom configurations. Our approach is “embarrassingly parallelisable” and capable of processing a point cloud containing over 100K points at eight orientations in less than 0.5s. For the object classes car, pedestrian and bicyclist the resulting detector achieves best-in-class detection and timing performance relative to prior art on the KITTI dataset as well as compared to another existing 3D object detection approach.

Subfields: Computational Mechanics, Computer Vision and Pattern Recognition, Environmental Engineering, Geology

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

In Computer Vision and Pattern Recognition

In Environmental Engineering

In Geology

Citations in the corpus

PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation (2016, methodology): “FPNN [12] and Vote3D [23] proposed special methods to deal with the sparsity problem; however, their operations are still on sparse volumes, it’s challenging for them to process very large point clouds.”
VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection (2017, background): “[43, 9] encode each nonempty voxel with 6 statistical quantities that are derived from all the points contained within the voxel.”
3D fully convolutional network for vehicle detection in point cloud (2016, background): “[16] also suggest it possible to predict 3D object localization by 2D CNN network on range scans.”
Multi-view 3D Object Detection Network for Autonomous Driving (2016, methodology): “Existing work usually encodes 3D LIDAR point cloud into a 3D grid [22, 5] or a front view map [14].”
Vote3Deep: Fast object detection in 3D point clouds using efficient convolutional neural networks (2016, methodology): “Inspired by [5], we propose to exploit feature-centric voting to build efficient CNNs to detect objects in point clouds natively in 3D – that is to say without projecting the input into a lower-dimensional space first or constraining the search space of the detector (Fig.”
Vehicle Detection from 3D Lidar Using Fully Convolutional Network (2016, methodology): “Wang and Posner [31] performs a sliding window based strategy to generate candidates and Behley et al. [2], Plotkin [22] segment the point cloud to generate detection candidates.”
A Survey of Autonomous Driving: Common Practices and Emerging Technologies (2019, methodology): “Shortly thereafter, a similar approach was used on point clouds created by lidars [163].”
Orientation-boosted Voxel Nets for 3D Object Recognition (2016, background): “Wang & Posner [36] were among the ﬁrst ones to utilize point clouds to obtain object proposals.”
PIXOR: Real-time 3D Object Detection from Point Clouds (2018, methodology): “As a result, typical systems [6, 37, 20] only run at 1-2 FPS.”
PointGrid: A Deep Network for 3D Shape Understanding (2018, methodology): “VoxNet and its variants [57, 35, 54, 6, 39, 32] discretizes the 3D bounding box into 3D occupancy grid, then applies 3D convolutions in a similar way as in 2D images.”
A General Pipeline for 3D Detection of Vehicles (2018, positive): “[15] proposed a different approach where the point cloud was converted into 3D feature grids and a 3D detection window was slid through the feature grids to identify vehicles.”
HDNET: Exploiting HD Maps for 3D Object Detection (2018, background): “Some detectors [13, 14, 15] search objects in 3D space densely via sliding window.”
PointASNL: Robust Point Clouds Processing Using Nonlocal Neural Networks With Adaptive Sampling (2020, background): “, autonomous driving [29], robotics [37], and place recognition [23].”
3D Object Proposals Using Stereo Imagery for Accurate Object Class Detection (2016, neutral): “69 Vote3Dy [53] 56.”
MVX-Net: Multimodal VoxelNet for 3D Object Detection (2019, methodology): “Some of the techniques ([38], [9], [35], [35], [17]) represent 3D point cloud data using voxel occupancy grid representation, followed by the use of 3D convolutions to compute the 3D bounding boxes.”
3D Semantic Parsing of Large-Scale Indoor Spaces (2016, methodology): “Baselines: We compare our method against the top performing algorithms from the KITTI object detection [11] dataset, mBOW [6] and Vote3D [37].”