Point cloud labeling using 3D Convolutional Neural Network

By Jing Huang, Suya You. 2016

355 citations recorded by the source corpus.

Subfields: Computational Mechanics, 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 Environmental Engineering

In Geology

Citations in the corpus

OctNet: Learning Deep 3D Representations at High Resolutions (2016, methodology): “, for labeling 3D point clouds, inefficient slidingwindow techniques with a limited receptive field must be adopted [20].”
Dynamic Edge-Conditioned Filters in Convolutional Neural Networks on Graphs (2017, neutral): “Due to their natural irregularity and sparsity, so far the only way of processing point clouds using deep learning has been to first voxelize them before feeding them to a 3D CNN, be it for classification [25] or segmentation [18] purposes.”
OctNetFusion: Learning Depth Fusion from Data (2017, background): “By utilizing large datasets of 3D models and high capacity 3D convolutional neural networks (3D CNNs), our approach learns to smooth out sensor noise, deals with outliers and completes missing 3D geometry.”
Deep Learning for LiDAR Point Clouds in Autonomous Driving: A Review (2020, background): “At each scale, a neighboring cubic with selected length is constructed for a given point [112].”
FoldingNet: Point Cloud Auto-Encoder via Deep Grid Deformation (2017, background): “Applications of learning on point clouds include shape completion and recognition [57], unmanned autonomous vehicles [36], 3D object detection, recognition and classification [9, 33, 40, 41, 48, 49, 53], contour detection [21], layout inference [18], scene labeling [31], category discovery [60], poi”
Deep Learning for 3D Point Clouds: A Survey (2019, methodology): “[195] first divided TABLE 4 Comparative 3D Object Detection Results on the KITTI Test BEV Detection Benchmark”
Deep Projective 3D Semantic Segmentation (2017, background): “Despite its success for image data, the application of CNNs to 3D point cloud data [9,20,27] have been severely hindered due to several important factors.”
Revisiting Point Cloud Shape Classification with a Simple and Effective Baseline (2021, methodology): “…using Rendered Images and Voxels: Many works use images rendered from object meshes for 3D shape analysis (Maturana & Scherer, 2015; Wu et al., 2015b; Yu et al., 2018; Guo et al., 2016; Shi et al., 2015; Hackel et al., 2017; Song & Xiao, 2016; 2014; Huang & You, 2016; Tchapmi et al., 2017).”
A Review on Deep Learning Techniques Applied to Semantic Segmentation (2017, methodology): “…SharpMask[84] DeepMask Top-downreﬁnementmodule MultiPathNet[85] FastR-CNN+DeepMask Multipathinformationﬂowthroughnetwork Huang-3DCNN[86] Own3DCNN 3DCNNforvoxelizedpointclouds PointNet[87] OwnMLP-based Segmentationofunorderedpointsets ClockworkConvnet[88] FCN…”
Recurrent Slice Networks for 3D Segmentation of Point Clouds (2018, background): “[33, 16, 22, 12, 13] made the early attempts of applying end-to-end deep learning algorithms for 3D data analysis, including 3D shape recognition, 3D urban scene segmentation [13].”
PPFNet: Global Context Aware Local Features for Robust 3D Point Matching (2018, background): “This idea is quickly extended to more informative encoding such as TDF [40], TSDF [48] (Truncated Signed Distance Field), multi-label occupancy [45] and other different ones [20, 47].”
A survey on deep learning techniques for image and video semantic segmentation (2018, methodology): “Targets Name and Reference Architecture Accuracy Efficiency Training Instance Sequences Multi-modal 3D Source Code Contribution(s) Fully Convolutional Network[68] VGG-16(FCN) ✗ ✗ ✗ ✗ ✓ Forerunner U-Net[69] Fully CNN, 4 downsampling/upsampling steps ✗ ✗ ✗ ✗ ✓ Data augmentation, Skip-layer, Patch wise”
SEGCloud: Semantic Segmentation of 3D Point Clouds (2017, background): “tation. It works well for indoor semantic scene understanding, although there is no evidence that it scales to larger input dimensions without additional training or adaptation required. Huang et al. [32] present a 3D-FCNN for 3D semantic segmentation which produces coarse voxellevel segmentation. D”
SemanticKITTI: A Dataset for Semantic Scene Understanding of LiDAR Sequences (2019, methodology): “Voxel-based methods transforming the point cloud into a voxel-grid and then applying convolutional neural networks (CNN) with 3D convolutions for object classification [34] and semantic segmentation [26] were among the first investigated models, since they allowed to exploit architectures and insigh”
Multiresolution Tree Networks for 3D Point Cloud Processing (2018, background): “These include multiview representations that render a shape from a collection of views [31,39,37] or voxel-based representations [44,19,30,4,43] that discretize point occupancy onto a 3D grid.”
Computer Vision for Autonomous Vehicles: Problems, Datasets and State-of-the-Art (2017, background): “Huang & You (2016) propose a framework for labeling 3D point cloud data using a 3D Convolutional Neural Network (3D-CNN).”