LatticeNet: Fast Point Cloud Segmentation Using Permutohedral Lattices

Radu Alexandru Rosu, Peer Schütt, Jan Quenzel, Sven Behnke
12/12/2019
cs.CVcs.LGeess.IV

Official code: AIS-Bonn/lattice_net (48 stars, MIT, last commit 2023-03-20).

Abstract

Deep convolutional neural networks (CNNs) have shown outstanding performance in the task of semantically segmenting images. However, applying the same methods on 3D data still poses challenges due to the heavy memory requirements and the lack of structured data. Here, we propose LatticeNet, a novel approach for 3D semantic segmentation, which takes as input raw point clouds. A PointNet describes the local geometry which we embed into a sparse permutohedral lattice. The lattice allows for fast convolutions while keeping a low memory footprint. Further, we introduce DeformSlice, a novel learned data-dependent interpolation for projecting lattice features back onto the point cloud. We present results of 3D segmentation on various datasets where our method achieves state-of-the-art performance.

View on arXivView PDF

Code Implementations(1)

AIS-Bonn/lattice_netOfficial100%
MIT

Fast Point Cloud Segmentation Using Permutohedral Lattices

482C, C++May 26, 20203 years agoMIT

Similar Papers

Pixel Difference Convolutional Network for RGB-D Semantic Segmentation

86% match2023

Pointwise Convolutional Neural Networks

85% match1 repo1322017

3D Semantic Segmentation with Submanifold Sparse Convolutional Networks

84% match1 repo2.1k2017

Depth-aware CNN for RGB-D Segmentation

84% match1 repo3062018

RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds

84% match1 repo1.6k2019

ESPNet: Efficient Spatial Pyramid of Dilated Convolutions for Semantic Segmentation

84% match1 repo5422018

Cite this paper

@article{rosu2019latticenet,
  title  = {LatticeNet: Fast Point Cloud Segmentation Using Permutohedral Lattices},
  author = {Radu Alexandru Rosu and Peer Schütt and Jan Quenzel and Sven Behnke},
  year   = {2019},
  eprint = {1912.05905},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1912.05905}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1912.05905.svg)](https://codeofpaper.com/paper/1912.05905)

Discussion