Self-Supervised Pre-Training Boosts Semantic Scene Segmentation on LiDAR Data

Mariona Carós, Ariadna Just, Santi Seguí, Jordi Vitrià
9/5/2023
cs.CVcs.AI

No official code found. Best implementation: marionacaros/barlow-twins-for-sem-seg (4 stars, MIT, last commit 2024-06-14).

Abstract

Airborne LiDAR systems have the capability to capture the Earth's surface by generating extensive point cloud data comprised of points mainly defined by 3D coordinates. However, labeling such points for supervised learning tasks is time-consuming. As a result, there is a need to investigate techniques that can learn from unlabeled data to significantly reduce the number of annotated samples. In this work, we propose to train a self-supervised encoder with Barlow Twins and use it as a pre-trained network in the task of semantic scene segmentation. The experimental results demonstrate that our unsupervised pre-training boosts performance once fine-tuned on the supervised task, especially for under-represented categories.

View on arXivView PDF

Code Implementations(1)

marionacaros/barlow-twins-for-sem-segCommunity100%
MIT

Self-Supervised Pre-Training Boosts Semantic Scene Segmentation on LiDAR data (MVA2023)

41PythonJun 30, 20232 years agoMIT
lidarssl

Similar Papers

Contrastive Learning for Self-Supervised Pre-Training of Point Cloud Segmentation Networks With Image Data

88% match2023

Understanding Self-Supervised Features for Learning Unsupervised Instance Segmentation

86% match2023

Bayesian Self-Training for Semi-Supervised 3D Segmentation

86% match2024

Semantic-Free Procedural 3D Shapes Are Surprisingly Good Teachers

86% match2024

You Only Need One Thing One Click: Self-Training for Weakly Supervised 3D Scene Understanding

85% match1 repo592023

CLIP2Scene: Towards Label-efficient 3D Scene Understanding by CLIP

85% match1 repo1872023

Cite this paper

@article{cars2023selfsupervised,
  title  = {Self-Supervised Pre-Training Boosts Semantic Scene Segmentation on LiDAR Data},
  author = {Mariona Carós and Ariadna Just and Santi Seguí and Jordi Vitrià},
  year   = {2023},
  eprint = {2309.02139},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2309.02139v2}
}

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/2309.02139.svg)](https://codeofpaper.com/paper/2309.02139)

Discussion