# Learning Representations and Generative Models for 3D Point Clouds

arXiv: https://arxiv.org/abs/1707.02392 · HTML: https://mlsanity.com/paper/1707.02392/

## Key facts

- Cited by 21 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 8 Jul 2017.

## Built on by (most cited first, 21 of 21)

- [Spatio-temporal Self-Supervised Representation Learning for 3D Point Clouds](https://mlsanity.com/paper/2109.00179/) (2021)
- [Point-JEPA: A Joint Embedding Predictive Architecture for Self-Supervised Learning on Point Cloud](https://mlsanity.com/paper/2404.16432/) (2024)
- [Learning Compositional Representation for 4D Captures with Neural ODE](https://mlsanity.com/paper/2103.08271/) (2021)
- [3DIAS: 3D Shape Reconstruction with Implicit Algebraic Surfaces](https://mlsanity.com/paper/2108.08653/) (2021)
- [Neural Vector Fields: Implicit Representation by Explicit Learning](https://mlsanity.com/paper/2303.04341/) (2023)
- [H4D: Human 4D Modeling by Learning Neural Compositional Representation](https://mlsanity.com/paper/2203.01247/) (2022)
- [Deep Hybrid Self-Prior for Full 3D Mesh Generation](https://mlsanity.com/paper/2108.08017/) (2021)
- [Object-Centric Learning with Slot Mixture Module](https://mlsanity.com/paper/2311.04640/) (2023)
- [LoRD: Local 4D Implicit Representation for High-Fidelity Dynamic Human Modeling](https://mlsanity.com/paper/2208.08622/) (2022)
- [Poisoning MorphNet for Clean-Label Backdoor Attack to Point Clouds](https://mlsanity.com/paper/2105.04839/) (2021)
- [DSLR: Dynamic to Static LiDAR Scan Reconstruction Using Adversarially Trained Autoencoder](https://mlsanity.com/paper/2105.12774/) (2021)
- [Explaining Deep Neural Networks for Point Clouds using Gradient-based Visualisations](https://mlsanity.com/paper/2207.12984/) (2022)
- [Concept Activation Vectors for Generating User-Defined 3D Shapes](https://mlsanity.com/paper/2205.02102/) (2022)
- [Neural Vector Fields: Generalizing Distance Vector Fields by Codebooks and Zero-Curl Regularization](https://mlsanity.com/paper/2309.01512/) (2023)
- [Points2Vec: Unsupervised Object-level Feature Learning from Point Clouds](https://mlsanity.com/paper/2102.04136/) (2021)
- [3D Intracranial Aneurysm Classification and Segmentation via Unsupervised Dual-branch Learning](https://mlsanity.com/paper/2201.02198/) (2022)
- [PointACL:Adversarial Contrastive Learning for Robust Point Clouds Representation under Adversarial Attack](https://mlsanity.com/paper/2209.06971/) (2022)
- [SLACK: Attacking LiDAR-based SLAM with Adversarial Point Injections](https://mlsanity.com/paper/2504.03089/) (2025)
- [Point2Point : A Framework for Efficient Deep Learning on Hilbert sorted Point Clouds with applications in Spatio-Temporal Occupancy Prediction](https://mlsanity.com/paper/2306.16306/) (2023)
- [Unlocking Hidden Potential in Point Cloud Networks with Attention-Guided Grouping-Feature Coordination](https://mlsanity.com/paper/2509.16639/) (2025)
- [Electrostatics-Inspired Surface Reconstruction (EISR): Recovering 3D Shapes as a Superposition of Poisson's PDE Solutions](https://mlsanity.com/paper/2602.11642/) (2026)

## Abstract (arXiv)

Three-dimensional geometric data offer an excellent domain for studying representation learning and generative modeling. In this paper, we look at geometric data represented as point clouds. We introduce a deep AutoEncoder (AE) network with state-of-the-art reconstruction quality and generalization ability. The learned representations outperform existing methods on 3D recognition tasks and enable shape editing via simple algebraic manipulations, such as semantic part editing, shape analogies and shape interpolation, as well as shape completion. We perform a thorough study of different generative models including GANs operating on the raw point clouds, significantly improved GANs trained in the fixed latent space of our AEs, and Gaussian Mixture Models (GMMs). To quantitatively evaluate generative models we introduce measures of sample fidelity and diversity based on matchings between sets of point clouds. Interestingly, our evaluation of generalization, fidelity and diversity reveals that GMMs trained in the latent space of our AEs yield the best results overall.