# Multi-Resolution Graph Neural Network for Large-Scale Pointcloud Segmentation

Source: [Multi-Resolution Graph Neural Network for Large-Scale Pointcloud Segmentation](<https://www.sota2.com/research/paper/68fa692e1e209131cea209c2>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-01-04T04:15:24.879Z

Authors: Liuyue Xie, Tomotake Furuhata, Kenji Shimada

Published: 2020-09-18T16:42:02.000Z

## Abstract

In this paper, we propose a multi-resolution deep-learning architecture to semantically segment dense large-scale pointclouds. Dense pointcloud data require a computationally expensive feature encoding process before semantic segmentation. Previous work has used different approaches to drastically downsample from the original pointcloud so common computing hardware can be utilized. While these approaches can relieve the computation burden to some extent, they are still limited in their processing capability for multiple scans. We present MuGNet, a memory-efficient, end-to-end graph neural network framework to perform semantic segmentation on large-scale pointclouds. We reduce the computation demand by utilizing a graph neural network on the preformed pointcloud graphs and retain the precision of the segmentation with a bidirectional network that fuses feature embedding at different resolutions. Our framework has been validated on benchmark datasets including Stanford Large-Scale 3D Indoor Spaces Dataset(S3DIS) and Virtual KITTI Dataset. We demonstrate that our framework can process up to 45 room scans at once on a single 11 GB GPU while still surpassing other graph-based solutions for segmentation on S3DIS with an 88.5\\% (+3\\%) overall accuracy and 69.8\\% (+7.7\\%) mIOU accuracy.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [3D Semantic Segmentation on S3DIS (all six areas)](<https://www.sota2.com/research/sota/3d-semantic-segmentation-on-s3dis-all-six-areas.md>): 3D Semantic Segmentation; S3DIS 1 (all six areas); Overall Accuracy (OA): 88.5.

- [3D Semantic Segmentation on vKITTI (6-fold val)](<https://www.sota2.com/research/sota/3d-semantic-segmentation-on-vkitti-6-fold-val.md>): 3D Semantic Segmentation; vKITTI (6-fold val); OA (Overall Accuracy): 85.1.

- [Semantic Segmentation on S3DIS (Area 5)](<https://www.sota2.com/research/sota/semantic-segmentation-on-s3dis-area-5.md>): Semantic segmentation; S3DIS (Area 5); mIOU: 63.5.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2009.08924>)

[PDF](<https://arxiv.org/pdf/2009.08924.pdf>)