LABELMAKER: Automatic Semantic Label Generation from RGB-D Trajectories

Silvan Weder, Hermann Blum, Francis Engelmann, Marc Pollefeys
11/20/2023
cs.CV

No official code found. Best implementation: cvg/LabelMaker (112 stars, BSD-3-Clause, last commit 2025-04-01).

Abstract

Semantic annotations are indispensable to train or evaluate perception models, yet very costly to acquire. This work introduces a fully automated 2D/3D labeling framework that, without any human intervention, can generate labels for RGB-D scans at equal (or better) level of accuracy than comparable manually annotated datasets such as ScanNet. Our approach is based on an ensemble of state-of-the-art segmentation models and 3D lifting through neural rendering. We demonstrate the effectiveness of our LabelMaker pipeline by generating significantly better labels for the ScanNet datasets and automatically labelling the previously unlabeled ARKitScenes dataset. Code and models are available at https://labelmaker.org

View on arXivView PDF

Code Implementations(1)

cvg/LabelMakerCommunity100%
BSD-3-Clause
1129Shell, PythonFeb 8, 20231 years agoBSD-3-Clause

Similar Papers

ALPS: An Auto-Labeling and Pre-training Scheme for Remote Sensing Segmentation With Segment Anything Model

85% match1 repo212024

From CAD models to soft point cloud labels: An automatic annotation pipeline for cheaply supervised 3D semantic segmentation

84% match2023

Thinking Like an Annotator: Generation of Dataset Labeling Instructions

84% match2023

Label Filling via Mixed Supervision for Medical Image Segmentation from Noisy Annotations

83% match2024

AllSpark: Reborn Labeled Features from Unlabeled in Transformer for Semi-Supervised Semantic Segmentation

83% match1 repo902024

A Unified Framework for Semi-Supervised Image Segmentation and Registration

83% match1 repo22025

Cite this paper

@article{weder2023labelmaker,
  title  = {LABELMAKER: Automatic Semantic Label Generation from RGB-D Trajectories},
  author = {Silvan Weder and Hermann Blum and Francis Engelmann and Marc Pollefeys},
  year   = {2023},
  eprint = {2311.12174},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2311.12174v1}
}

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/2311.12174.svg)](https://codeofpaper.com/paper/2311.12174)

Discussion