3D Representation Learning

Overview

3D representation learning studies how to encode point clouds, meshes, voxel grids, implicit fields, Gaussian splats, and multi-view images into features that preserve shape, spatial relations, and meaning. Methods include masked and contrastive self-supervised learning, rotation-equivariant networks, neural fields, generative latent models, and alignment with image, text, or action features. Systems are tested on tasks such as ShapeNet object recognition and reconstruction, ScanNet scene understanding, and robotic manipulation benchmarks. A central challenge is balancing compact, efficient features with the geometric detail, topology, and robustness needed for sparse views, occlusion, and transfer across datasets or tasks.

Latest papers

CardsList

207 papers16 papers in the last four weeks, down 27% on the four weeks before. 0.1% of all new papers.

  1. Do 3D Point Representations Encode Architectural Functions? Probing Supervised and Pretrained Models on Urban Facades

    Nov 18, 2026Abdiaziz Muse, Jiju PoovvancheriPoint Cloud Understanding3D Representation Learning

  2. BRepA: Graph-Based B-Rep Pretraining for Unified CAD Understanding and Generation

    Nov 17, 2026Dekai Zhu, Zhaiyu Chen, Stefan Gavranovic +13D Representation LearningSelf-Supervised Pre-Training

  3. CascadeJEPA: Enhancing 3D Geometric Representations via Spatially Coupled Cascaded Prediction

    Nov 17, 2026Kailai Zhou, Ziqi Chen, Xiaoning Xu +8Joint-Embedding Predictive Architecture3D Representation Learning

  4. Lift, See, Act: Hierarchical Robot Policy Pretraining with 3D Foundation Models

    Nov 16, 2026Yiyuan Ge, Changxing Ding, Ziyu Hao +2Robot LearningRobot Learning from Video

  5. Split-and-scale Latent 3D Representations

    Nov 12, 2026Dorian Chan, Muhammed Kocabas, Xiaoming Zhao +23D Scene Representation3D Representation Learning

  6. Rethinking Radar 3D Perception: Representation Learning from Raw Spectra

    Nov 9, 2026Miao Zhang, Mohamed Abdelsamad, Bin Yang +5Representation Learning3D Representation Learning

  7. SurfScaff3D: Bridging Geometric and Semantic Representations for LLM-based 3D Object Generation

    Nov 8, 2026Anay Loya, Rishi Gupta, Arnav Samal +1Text-to-3D GenerationPart-Level 3D Generation

  8. Spatial Representation Distillation and Knowledge Routing for Vision-Language-Action Models

    Nov 7, 2026Jungin Park, Chaoran Zhu, Changjae Oh +1Vision-Language-Action ModelsEfficient VLA Models

  9. PoseBench: Probing Pose Awareness in Visual Foundation Models

    Nov 7, 2026Most Nilufa Yeasmin, Draix Wyatt, Ruyi LianVision Foundation ModelsVisual Representation Learning

  10. ZeoBench: A Benchmark for Self-Supervised Learning on 3D Zeolite Representations

    Nov 6, 2026Aaron Sun, Yachan Liu, Ping Yang +23D Representation LearningSelf-Supervised Pre-Training

  11. Aligning Policy Encoders with a Shape Teacher

    Nov 5, 2026Madelon Bernardy, Mohammad Mahdi Derakhshani, David M Knigge +2Visuomotor Policy LearningRobotic Manipulation

  12. VINO: Revisiting 3D Representation Learning with Volume Transformers

    Nov 4, 2026Adrian Kruse, Kadir Yilmaz, Daan de Geus +23D Representation LearningRepresentation Learning

  13. Lie Routing Vision Language Models for View-Robust 3D Understanding

    Nov 1, 2026Jingjun Yi, Chen Hu, Qi Bi +73D VLMs3D Scene Understanding

  14. Rethinking Asymmetry in Self-Supervised 3D Geometry Learning

    Oct 30, 2026Gyeongjin Kang, Freeman Cheng, Botao Ye +33D Representation LearningMulti-View 3D Reconstruction

  15. PT3^3: Prototype-Guided Self-Supervised Pretraining for Generalizable mmWave Human Point Cloud Representation

    Oct 28, 2026Ziyang Liu, Kieren Yu, Xiao Yang +2Self-Supervised Point Cloud LearningmmWave Radar Sensing

  16. PhysCreator: Creating Physically-Attributed 3D Assets at Will

    Oct 25, 2026Shiyan Zheng, Ming Li, Lei Han +13D Asset Generation3D Representation Learning

  17. RangeNav: Task-Aligned LiDAR Range-Image Pretraining for Robot Navigation

    Oct 21, 2026Samuel Foo Enze, Manthan Patel, Fan Yang +4Robot Navigation3D Representation Learning

  18. Noise-Invariant Spatial Encoding for 3D Manipulation Policies under Depth Corruption

    Oct 17, 2026Zhiyuan Yang, Jiaju He, Ruiteng Zhao +3Robotic ManipulationNoise Robustness

  19. TriFold: Folding irregular 3D Gaussians into structured latents for 3D Scene Generation

    Oct 16, 2026Yun Wang, Chiyu Wang, Fan Jiang +53D Scene Generation3D Representation Learning

  20. Teaching DINOv3 About Partial 3D Geometry: A Self-Supervised Geometry-Aware Approach

    Oct 15, 2026Viktoria Ehm, Dongliang Cao, Riccardo Marin +43D Representation LearningSelf-Supervised Visual Representation Learning

  21. Learning Topology-Aware Representations for Mechanics-Relevant TPMS Metamaterials

    Oct 14, 2026Songlin Duan, Shuo Yang, Yujia Zheng +1Topological Representation Learning3D Representation Learning

  22. Q-MAGIC: Quad-only Mesh Autoregressive Generation via Singularity-Aware Circular Relative Offset Tokenization

    Oct 14, 2026Yuming Zhao, Zezeng Li, Weimin Wang +23D Mesh Generation3D Representation Learning

  23. Revisiting Triplet Contrastive Learning for 3D Shape Understanding

    Oct 14, 2026Chaoyang Zhu, Zhe Liao, Jiayi Zhang +33D Representation LearningCross-Modal Alignment

  24. Align-VAE: Aligning Attention and Geometry in VecSet VAEs for 3D Reconstruction

    Oct 10, 2026Huaizhi Qu, Mingxin Yang, Xintong Han +2Variational Autoencoders3D Reconstruction

  25. Point-Focused Attention Meets Context-Scan State Space: Robust Biological Visual Perception for Point Cloud Representation

    Oct 8, 2026Kanglin Qu, Pan Gao, Qun Dai +1Point Cloud Learning3D Representation Learning

  26. ActiveLang: Active Open-Vocabulary 3D Mapping with Semantic-Uncertainty-Guided Exploration

    Oct 7, 2026Liyan Chen, Hairong Yin, Huangying Zhan +33D Semantic Mapping3D Mapping