SPLATNet: Sparse Lattice Networks for Point Cloud Processing

@article{Su2018SPLATNetSL,
  title={SPLATNet: Sparse Lattice Networks for Point Cloud Processing},
  author={Hang Su and V. Jampani and Deqing Sun and Subhransu Maji and Evangelos Kalogerakis and Ming-Hsuan Yang and Jan Kautz},
  journal={2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year={2018},
  pages={2530-2539}
}
  • Hang Su, V. Jampani, +4 authors J. Kautz
  • Published 2018
  • Computer Science
  • 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition
We present a network architecture for processing point clouds that directly operates on a collection of points represented as a sparse set of samples in a high-dimensional lattice. [...] Key Method These layers maintain efficiency by using indexing structures to apply convolutions only on occupied parts of the lattice, and allow flexible specifications of the lattice structure enabling hierarchical and spatially-aware feature learning, as well as joint 2D-3D reasoning.Expand
LatticeNet: Fast Spatio-Temporal Point Cloud Segmentation Using Permutohedral Lattices
TLDR
This work proposes LatticeNet, a novel approach for 3D semantic segmentation, which takes raw point clouds as input and introduces DeformSlice, a novels learned data-dependent interpolation for projecting lattice features back onto the point cloud. Expand
Point Clouds Learning with Attention-based Graph Convolution Networks
TLDR
This work proposes a novel network structure, named Attention-based Graph Convolution Networks (AGCN), to extract point clouds features, taking the learning process as a message propagation between adjacent points for analyzing the relationships between local features of the points. Expand
ConvPoint: continuous convolutions for cloud processing
TLDR
This paper proposes a generalization of the discrete convolutional neural networks able to deal with sparse input point cloud and replaces the discrete kernels by continuous ones, which can easily be used for neural network design similarly to 2D CNNs. Expand
Point Cloud Processing via Recurrent Set Encoding
TLDR
A new permutation-invariant network for 3D point cloud processing that is effective at spatial feature learning, and competes favorably with the state-of-the-arts (SOTAs) on a number of benchmarks, and significantly more efficient compared to the SOTAs. Expand
Multiresolution Tree Networks for 3D Point Cloud Processing
TLDR
This model represents a 3D shape as a set of locality-preserving 1D ordered list of points at multiple resolutions, which allows efficient feed-forward processing through 1D convolutions, coarse-to-fine analysis through a multi-grid architecture, and it leads to faster convergence and small memory footprint during training. Expand
HPGCNN: Hierarchical Parallel Group Convolutional Neural Networks for Point Clouds Processing
TLDR
This work introduces HPGCNN, an efficient and lightweight neural architecture that clearly surpasses state-of-the-art approaches for point clouds classification dataset ModelNet40 and large scale semantic segmentation datasets ShapeNet Parts, S3DIS, vKITTI and SemanticKittI in terms of accuracy and complexity. Expand
Differentiable Convolution Search for Point Cloud Processing
TLDR
PointSeaConv is proposed, a novel differential convolution search paradigm on point clouds that can work in a purely data-driven manner and thus is capable of auto-creating a group of suitable convolutions for geometric shape modeling and introduces epsilon-greedy algorithm to alleviate the effect of discretization error. Expand
DRINet: A Dual-Representation Iterative Learning Network for Point Cloud Segmentation
TLDR
The DRINet work, which serves as the basic network structure for dual-representation learning with great flexibility at feature transferring and less computation cost, especially for large-scale point clouds, is proposed. Expand
Point Attention Network for Semantic Segmentation of 3D Point Clouds
TLDR
The proposed point attention network consists of an encoder and decoder which, together with the LAE-Conv layers and the point-wise spatial attention modules, make it an end-to-end trainable network for predicting dense labels for 3D point cloud segmentation. Expand
Interpolated Convolutional Networks for 3D Point Cloud Understanding
TLDR
A novel Interpolated Convolution operation, InterpConv, is proposed to tackle the point cloud feature learning and understanding problem, to utilize a set of discrete kernel weights and interpolate point features to neighboring kernel-weight coordinates by an interpolation function for convolution. Expand
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 62 REFERENCES
PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation
TLDR
This paper designs a novel type of neural network that directly consumes point clouds, which well respects the permutation invariance of points in the input and provides a unified architecture for applications ranging from object classification, part segmentation, to scene semantic parsing. Expand
Escape from Cells: Deep Kd-Networks for the Recognition of 3D Point Cloud Models
TLDR
A new deep learning architecture that is designed for 3D model recognition tasks and works with unstructured point clouds, Kd-networks, which demonstrates competitive performance in a number of shape recognition tasks such as shape classification, shape retrieval and shape part segmentation. Expand
Deep Learning on Point Sets for 3 D Classification and Segmentation
Point cloud is an important type of geometric data structure. Due to its irregular format, most researchers transform such data to regular 3D voxel grids or collections of images. This, however,Expand
GWCNN: A Metric Alignment Layer for Deep Shape Analysis
TLDR
This work proposes to optimize for the geometric representation during the network learning process using a novel metric alignment layer that maps unstructured geometric data to a regular domain by minimizing the metric distortion of the map using the regularized Gromov–Wasserstein objective. Expand
PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space
TLDR
A hierarchical neural network that applies PointNet recursively on a nested partitioning of the input point set and proposes novel set learning layers to adaptively combine features from multiple scales to learn deep point set features efficiently and robustly. Expand
Submanifold Sparse Convolutional Networks
TLDR
This work introduces a sparse convolutional operation tailored to processing sparse data that operates strictly on submanifolds, rather than "dilating" the observation with every layer in the network. Expand
Hypercolumns for object segmentation and fine-grained localization
TLDR
Using hypercolumns as pixel descriptors, this work defines the hypercolumn at a pixel as the vector of activations of all CNN units above that pixel, and shows results on three fine-grained localization tasks: simultaneous detection and segmentation, and keypoint localization. Expand
OctNet: Learning Deep 3D Representations at High Resolutions
TLDR
The utility of the OctNet representation is demonstrated by analyzing the impact of resolution on several 3D tasks including 3D object classification, orientation estimation and point cloud labeling. Expand
Octree Generating Networks: Efficient Convolutional Architectures for High-resolution 3D Outputs
We present a deep convolutional decoder architecture that can generate volumetric 3D outputs in a compute- and memory-efficient manner by using an octree representation. The network learns to predictExpand
OctNetFusion: Learning Depth Fusion from Data
TLDR
This paper presents a novel 3D CNN architecture that learns to predict an implicit surface representation from the input depth maps and significantly outperforms the traditional volumetric fusion approach in terms of noise reduction and outlier suppression. Expand
...
1
2
3
4
5
...