PointGrid: A Deep Network for 3D Shape Understanding

By Truc Le, Ye Duan. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018

408 citations recorded by the source corpus.

Volumetric grid is widely used for 3D deep learning due to its regularity. However the use of relatively lower order local approximation functions such as piece-wise constant function (occupancy grid) or piece-wise linear function (distance field) to approximate 3D shape means that it needs a very high-resolution grid to represent finer geometry details, which could be memory and computationally inefficient. In this work, we propose the PointGrid, a 3D convolutional network that incorporates a constant number of points within each grid cell thus allowing the network to learn higher order local approximation functions that could better represent the local geometry shape details. With experiments on popular shape recognition benchmarks, PointGrid demonstrates state-of-the-art performance over existing deep learning methods on both classification and segmentation.

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

View Paper: https://www.semanticscholar.org/paper/e81acd4f35eb2202bf753266ec83d3174645820f

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 Geology

Citations in the corpus

ShellNet: Efficient Point Cloud Convolutional Neural Networks Using Concentric Shells Statistics (2019, methodology): “[17] propose to apply convolution on a regular grid with each cell containing point features that are resampled to a fixed size.”
PartNet: A Large-Scale Benchmark for Fine-Grained and Hierarchical Part-Level 3D Object Understanding (2018, background): “Recent advances of 3D semantic segmentation [30, 31, 46, 19, 35, 24, 9, 39, 40, 42, 33, 7, 26, 23] have accomplished promising performance in coarse-level segmentation on the ShapeNet Part dataset [3, 45].”
Modeling Point Clouds With Self-Attention and Gumbel Subset Sampling (2019, background): “, graph neural networks on point sets [46, 34, 23, 22], or 3) hybrid [55, 21, 36].”
3D Shape Generation and Completion through Point-Voxel Diffusion (2021, methodology): “Due to the correspondence between voxels and 2D pixels, many works have explored voxel-based classification and segmentation using volumetric convolution [26, 31, 19, 39, 42, 7].”
Point-Voxel CNN for Efficient 3D Deep Learning (2019, background): “Recent studies suggest that the volumetric representation can also be used in 3D shape segmentation [21, 37, 44] and 3D object detection [50].”
TopNet: Structural Point Cloud Decoder (2019, background): “Volumetric 3D shape completion: Earlier and some recent works on shape completion leveraging neural networks favored voxel grids, and distance fields representations [6, 15, 29, 27, 18] which are well suited for processing with 3D convolutional neural networks.”
PointASNL: Robust Point Clouds Processing Using Nonlocal Neural Networks With Adaptive Sampling (2020, neutral): “8 PointGrid [18] pnt 1k - 92.”
RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds (2019, background): “Alternatively, point clouds can be voxelized into 3D grids and then powerful 3D CNNs are applied in [14, 28, 10, 39, 9].”
PCT: Point cloud transformer (2020, neutral): “9% PointGrid [14] P 1k 92.”
Learning Object Bounding Boxes for 3D Instance Segmentation on Point Clouds (2019, neutral): “Recent learning based approaches mainly include voxel-based [42; 46; 41; 23; 40; 11; 4] and point-based schemes [37; 19; 14; 16; 45].”
GraspNet-1Billion: A Large-Scale Benchmark for General Object Grasping (2020, methodology): “After that, many methods [34, 38, 3, 23, 12, 43, 39, 40, 20, 19, 49, 47, 44, 46, 15] are proposed to perform point cloud classification and segmentation.”
Deep Hough Voting for 3D Object Detection in Point Clouds (2019, background): “Recently we see a surge of interest in designing deep network architectures suited for point clouds [35, 36, 43, 1, 25, 9, 48, 45, 46, 22, 17, 53, 52, 49, 51], which showed remarkable performance in 3D object classification, object part segmentation, as well as scene segmentation.”
Rotation Invariant Convolutions for 3D Point Clouds Deep Learning (2019, background): “A wide body of works [24, 26, 12, 17, 19, 35, 39, 3] have recently been proposed to solve this problem, demonstrating state-of-the-art performance in scene understanding with point cloud data.”