LatticeNet: Fast Point Cloud Segmentation Using Permutohedral Lattices

TL;DR

LatticeNet is a fast and memory-efficient 3D point cloud segmentation method that uses permutohedral lattices and a novel interpolation technique to achieve state-of-the-art results.

Contribution

It introduces LatticeNet, combining permutohedral lattices with learned interpolation for improved 3D semantic segmentation of raw point clouds.

Findings

Achieves state-of-the-art segmentation accuracy on multiple datasets.

Offers a fast convolution method with low memory usage.

Introduces DeformSlice, a novel learned interpolation technique.

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.