mlsanityData to 13 Oct 2026

Learnable Earth Parser: Discovering 3D Prototypes in Aerial Scans

arXiv 2304.09704 ↗First posted 19 Apr 2023cs.CVcited by 5

Not on the map. Why?

Abstract (arXiv)

We propose an unsupervised method for parsing large 3D scans of real-world scenes with easily-interpretable shapes. This work aims to provide a practical tool for analyzing 3D scenes in the context of aerial surveying and mapping, without the need for user annotations. Our approach is based on a probabilistic reconstruction model that decomposes an input 3D point cloud into a small set of learned prototypical 3D shapes. The resulting reconstruction is visually interpretable and can be used to perform unsupervised instance and low-shot semantic segmentation of complex scenes. We demonstrate the usefulness of our model on a novel dataset of seven large aerial LiDAR scans from diverse real-world scenarios. Our approach outperforms state-of-the-art unsupervised methods in terms of decomposition accuracy while remaining visually interpretable. Our code and dataset are available at https://romainloiseau.fr/learnable-earth-parser/

Key facts. Cited by 5 papers in the map (data through 13 Oct 2026). Builds on 35 of its 83 references that are in the map. First posted on arXiv 19 Apr 2023.

Builds on

35 of 83 references.

References · 35

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,093
  2. 2016Layer Normalizationcited by 5,624
  3. 2015ShapeNet: An Information-Rich 3D Model Repositorycited by 2,889
  4. 2020Object-Centric Learning with Slot Attentioncited by 869
  5. 2020Accelerating 3D Deep Learning with PyTorch3Dcited by 784
  6. 2021KITTI-360: A Novel Dataset and Benchmarks for Urban Scene Understanding in 2D and 3Dcited by 571
  7. 2018AtlasNet: A Papier-Mach'e Approach to Learning 3D Surface Generationcited by 539
  8. 2018Invariant Information Clustering for Unsupervised Image Classification and Segmentationcited by 455
  9. 2020A Survey on Neural Network Interpretabilitycited by 291
  10. 2021Self-Supervised Pretraining of 3D Features on any Point-Cloudcited by 217
  11. 2016Learning Shape Abstractions by Assembling Volumetric Primitivescited by 181
  12. 2019Superquadrics Revisited: Learning 3D Shape Parsing beyond Cuboidscited by 124
23 more
  1. 2017CSGNet: Neural Shape Parser for Constructive Solid Geometrycited by 106
  2. 20183D Point Capsule Networkscited by 106
  3. 2021Neural Parts: Learning Expressive 3D Shape Abstractions with Invertible Neural Networkscited by 103
  4. 2021Unsupervised Discovery of Object Radiance Fieldscited by 101
  5. 2019Learning elementary structures for 3D shape generation and matchingcited by 98
  6. 2018Supervised Fitting of Geometric Primitives to 3D Point Cloudscited by 91
  7. 2019Stacked Capsule Autoencoderscited by 82
  8. 20173D-PRNN: Generating Shape Primitives with Recurrent Neural Networkscited by 77
  9. 2020Learning Unsupervised Hierarchical Part Decomposition of 3D Objects from a Single RGB Imagecited by 72
  10. 2020DALES: A Large-scale Aerial LiDAR Data Set for Semantic Segmentationcited by 65
  11. 2021Unsupervised Layered Image Decomposition into Object Prototypescited by 50
  12. 2021MarioNette: Self-Supervised Sprite Learningcited by 35
  13. 2019DublinCity: Annotated LiDAR Point Cloud and its Applicationscited by 31
  14. 2021Robust and Accurate Superquadric Recovery: a Probabilistic Approachcited by 29
  15. 2020Deep Transformation-Invariant Clusteringcited by 26
  16. 2021SUM: A Benchmark Dataset of Semantic Urban Meshescited by 23
  17. 2022City3D: Large-Scale Building Reconstruction from Airborne LiDAR Point Cloudscited by 22
  18. 2020Joint Supervised and Self-Supervised Learning for 3D Real-World Challengescited by 21
  19. 2021Cuboids Revisited: Learning Robust 3D Shape Fitting to Single RGB Imagescited by 17
  20. 2022Primitive-based Shape Abstraction via Nonparametric Bayesian Inferencecited by 15
  21. 2022MonteBoxFinder: Detecting and Filtering Primitives to Fit a Noisy Point Cloudcited by 11
  22. 2021Representing Shape Collections with Alignment-Aware Linear Modelscited by 7
  23. 2022Multi-Layer Modeling of Dense Vegetation from Aerial LiDAR Scanscited by 6

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2023: 1; 2024: 2; 2025: 1; 2026: 1.

5 papers.

  1. 2023Differentiable Blocks World: Qualitative 3D Decomposition by Rendering Primitivescited by 38
  2. 2025ForestFormer3D: A Unified Framework for End-to-End Segmentation of Forest LiDAR 3D Point Cloudscited by 13
  3. 2024Self-supervised Learning of Hybrid Part-aware 3D Representations of 2D Gaussians and Superquadricscited by 8
  4. 2024Deep Learning on 3D Semantic Segmentation: A Detailed Reviewcited by 6
  5. 2026Deep Sprite-based Image Models: An Analysiscited by 0