mlsanityData to 7 Oct 2026

Native and Compact Structured Latents for 3D Generation

arXiv 2512.14692 ↗First posted 16 Dec 2025cs.CV cs.AIcited by 157

Not on the map. Why?

Cited by 25 in Diffusion models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Recent advancements in 3D generative modeling have significantly improved the generation realism, yet the field is still hampered by existing representations, which struggle to capture assets with complex topologies and detailed appearance. This paper present an approach for learning a structured latent representation from native 3D data to address this challenge. At its core is a new sparse voxel structure called O-Voxel, an omni-voxel representation that encodes both geometry and appearance. O-Voxel can robustly model arbitrary topology, including open, non-manifold, and fully-enclosed surfaces, while capturing comprehensive surface attributes beyond texture color, such as physically-based rendering parameters. Based on O-Voxel, we design a Sparse Compression VAE which provides a high spatial compression rate and a compact latent space. We train large-scale flow-matching models comprising 4B parameters for 3D generation using diverse public 3D asset datasets. Despite their scale, inference remains highly efficient. Meanwhile, the geometry and material quality of our generated assets far exceed those of existing models. We believe our approach offers a significant advancement in 3D generative modeling.

Key facts. Cited by 157 papers in the map (data through 7 Oct 2026). Builds on 68 of its 81 references that are in the map. First posted on arXiv 16 Dec 2025.

Builds on

68 of 81 references.

References · 68

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  3. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,141
  4. 2022Scalable Diffusion Models with Transformerscited by 5,931
  5. 2022Flow Matching for Generative Modelingcited by 5,426
  6. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,581
  7. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,094
  8. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,550
  9. 2022A ConvNet for the 2020scited by 3,545
  10. 2019DeepSDF: Learning Continuous Signed Distance Functions for Shape Representationcited by 2,682
  11. 2018Occupancy Networks: Learning 3D Reconstruction in Function Spacecited by 2,037
  12. 2022Objaverse: A Universe of Annotated 3D Objectscited by 1,522
38 more
  1. 2019Root Mean Square Layer Normalizationcited by 1,425
  2. 2025DINOv3cited by 1,297
  3. 2023PixArt-α: Fast Training of Diffusion Transformer for Photorealistic Text-to-Image Synthesiscited by 854
  4. 2023Objaverse-XL: A Universe of 10M+ 3D Objectscited by 767
  5. 2024Structured 3D Latents for Scalable and Versatile 3D Generationcited by 742
  6. 2021Diffusion Probabilistic Models for 3D Point Cloud Generationcited by 734
  7. 2022Point-E: A System for Generating 3D Point Clouds from Complex Promptscited by 694
  8. 2023Mip-Splatting: Alias-free 3D Gaussian Splattingcited by 661
  9. 2021Perceiver IO: A General Architecture for Structured Inputs & Outputscited by 616
  10. 2022LION: Latent Point Diffusion Models for 3D Shape Generationcited by 552
  11. 20213D Shape Generation and Completion through Point-Voxel Diffusioncited by 516
  12. 20233DShape2VecSet: A 3D Shape Representation for Neural Fields and Generative Diffusion Modelscited by 449
  13. 2024CLAY: A Controllable Large-scale Generative Model for Creating High-quality 3D Assetscited by 444
  14. 2023Shap-E: Generating Conditional 3D Implicit Functionscited by 379
  15. 2021Extracting Triangular 3D Models, Materials, and Lighting From Imagescited by 363
  16. 2021ABO: Dataset and Benchmarks for Real-World 3D Object Understandingcited by 360
  17. 20203D-FUTURE: 3D Furniture shape with TextUREcited by 306
  18. 2020PolyGen: An Autoregressive Generative Model of 3D Meshescited by 267
  19. 2023Uni3D: Exploring Unified 3D Representation at Scalecited by 231
  20. 2023Michelangelo: Conditional 3D Shape Generation based on Shape-Image-Text Aligned Latent Representationcited by 225
  21. 2024Direct3D: Scalable Image-to-3D Generation via 3D Latent Diffusion Transformercited by 210
  22. 2024Deep Compression Autoencoder for Efficient High-Resolution Diffusion Modelscited by 209
  23. 2022DiffRF: Rendering-Guided 3D Radiance Field Diffusioncited by 204
  24. 2025TripoSG: High-Fidelity 3D Shape Synthesis using Large-Scale Rectified Flow Modelscited by 204
  25. 2023ULIP-2: Towards Scalable Multimodal Pre-training for 3D Understandingcited by 199
  26. 2023XCube: Large-Scale 3D Generative Modeling using Sparse Voxel Hierarchiescited by 183
  27. 20233DGen: Triplane Latent Diffusion for Textured Mesh Generationcited by 181
  28. 2023Habitat Synthetic Scenes Dataset (HSSD-200): An Analysis of 3D Scene Scale and Realism Tradeoffs for ObjectGoal Navigationcited by 162
  29. 2025Hunyuan3D 2.1: From Images to High-Fidelity 3D Assets with Production-Ready PBR Materialcited by 157
  30. 2024MeshAnything: Artist-Created Mesh Generation with Autoregressive Transformerscited by 140
  31. 2023Flexible Isosurface Extraction for Gradient-Based Mesh Optimizationcited by 138
  32. 2022Neural Wavelet-domain Diffusion for 3D Shape Generationcited by 132
  33. 2023Locally Attentional SDF Diffusion for Controllable 3D Shape Generationcited by 129
  34. 2020Deformed Implicit Field: Modeling 3D Shapes with Learned Dense Correspondencecited by 119
  35. 2021Using Shape to Categorize: Low-Shot Learning with an Explicit Shape Biascited by 115
  36. 2025Direct3D-S2: Gigascale 3D Generation Made Easy with Spatial Sparse Attentioncited by 103
  37. 20243DTopia-XL: Scaling High-quality 3D Asset Generation via Primitive Diffusioncited by 100
  38. 2025Hi3DGen: High-fidelity 3D Geometry Generation from Images via Normal Bridgingcited by 95

50 most cited of 68 All 68 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 1; 2026: 156.

157 papers.

Sort:

By area:

  1. 2026SceneSmith: Agentic Generation of Simulation-Ready Indoor Scenescited by 39
  2. 2025UltraShape 1.0: High-Fidelity 3D Shape Generation via Scalable Geometric Refinementcited by 16
  3. 2026LATO: 3D Mesh Flow Matching with Structured TOpology Preserving LAtentscited by 10
  4. 2026SegviGen: Repurposing 3D Generative Model for Part Segmentationcited by 9
  5. 2026PhysX-Omni: Unified Simulation-Ready Physical 3D Generation for Rigid, Deformable, and Articulated Objectscited by 7
  6. 2026Stream3D: Sequential Multi-View 3D Generation via Evidential Memorycited by 6
  7. 2026LATO.2: Factorized 3D Mesh Generation with Vertex and Topology Flowcited by 6
  8. 2026PEGAsus: 3D Personalization of Geometry and Appearancecited by 5
  9. 2026PAct: Part-Decomposed Single-View Articulated Object Generationcited by 5
  10. 2026LottieGPT: Tokenizing Vector Animation for Autoregressive Generationcited by 5
  11. 2026DexWorldModel: Causal Latent World Modeling towards Automated Learning of Embodied Taskscited by 5
  12. 2026Reconstruction by Generation: 3D Multi-Object Scene Reconstruction from Sparse Observationscited by 5
38 more
  1. 2026Optimization and Generation in Aerodynamics Inverse Designcited by 4
  2. 2026GaussianGPT: Towards Autoregressive 3D Gaussian Scene Generationcited by 4
  3. 2026Seed3D 2.0: Advancing High-Fidelity Simulation-Ready 3D Content Generationcited by 4
  4. 2026Generative 3D Gaussians with Learned Density Controlcited by 4
  5. 2026EmbodiedGen V2: An Agentic, Simulation-Ready 3D World Engine for Embodied AIcited by 4
  6. 2026MoGe-3: Fine-Detail Monocular Geometry Estimation with Self-Guided Sparse Volumetric Refinementcited by 4
  7. 2026Research on World Models Is Not Merely Injecting World Knowledge into Specific Taskscited by 3
  8. 2026Fast-SAM3D: 3Dfy Anything in Images but Fastercited by 3
  9. 2026VideoMatGen: PBR Materials through Joint Generative Modelingcited by 3
  10. 2026Omni123: Exploring 3D Native Foundation Models with Limited 3D Data by Unifying Text to 2D and 3D Generationcited by 3
  11. 2026AniGen: Unified S3 Fields for Animatable 3D Asset Generationcited by 3
  12. 2026Mix3R: Mixing Feed-forward Reconstruction and Generative 3D Priors for Joint Multi-view Aligned 3D Reconstruction and Pose Estimationcited by 3
  13. 2026GenRecon: Bridging Generative Priors for Multi-View 3D Scene Reconstructioncited by 3
  14. 2026LabVLA: Grounding Vision-Language-Action Models in Scientific Laboratoriescited by 3
  15. 2026World Tracing: Generative Pixel-Aligned Geometry Beyond the Visiblecited by 3
  16. 2026φ-Scene: Physically Grounded Image-to-3D Scene Reconstructioncited by 3
  17. 2026Meshy T2: Fast Native Mesh Generation with Flow Matchingcited by 3
  18. 2026LSRM: High-Fidelity Object-Centric Reconstruction via Scaled Context Windowscited by 2
  19. 2026Strips as Tokens: Artist Mesh Generation with Native UV Segmentationcited by 2
  20. 2026Hitem3D 2.0: Multi-View Guided Native 3D Texture Generationcited by 2
  21. 2026Any 3D Scene is Worth 1K Tokens: 3D-Grounded Representation for Scene Generation at Scalecited by 2
  22. 2026Asset Harvester: Extracting 3D Assets from Autonomous Driving Logs for Simulationcited by 2
  23. 2026DVD: Discrete Voxel Diffusion for 3D Generation and Editingcited by 2
  24. 2026EVA01: Unified Native 3D Understanding and Generation via Mixture-of-Transformerscited by 2
  25. 2026ROAR-3D: Routing Arbitrary Views for High-Fidelity 3D Generationcited by 2
  26. 2026AssetGen: Deployable 3D Asset Generation at Interactive Speedcited by 2
  27. 2026ABot-Earth 0.5: Generative 3D Earth Modelcited by 2
  28. 2026Instruct-Particulate: Scaling Feed-Forward 3D Object Articulation with Kinematic Controlcited by 2
  29. 2026TriFlow: Generating Artist-Like 3D Mesh Topology via Nearest-Vertex Vector Fieldscited by 2
  30. 2026Sat2City v2: Native 3D City Asset Generation from a Single Satellite Imagecited by 2
  31. 2026One Video, One World: Turning Monocular Video into Physical 4D Scenescited by 2
  32. 2026ShellMaker: Language-Guided Exterior Completion under Structural Constraintscited by 2
  33. 2026PixGS: Pixel-Space Diffusion for Direct 3D Gaussian Splat Generationcited by 2
  34. 2026Instant NuRec: Feed-Forward 3D Gaussian Reconstruction for Driving Scene Simulationcited by 2
  35. 2026Axolotl3D: a Unified Framework for Faithful 3D Shape Completioncited by 2
  36. 2026WorldSculpt: Generating Compositional Worlds from Grounded Videoscited by 2
  37. 2026NuiWorld: Exploring a Scalable Framework for End-to-End Controllable World Generationcited by 1
  38. 2026SK-Adapter: Skeleton-Based Structural Control for Native 3D Generationcited by 1

50 most cited of 157 All 157 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025LATTICE: Democratize High-Fidelity 3D Generation at Scale
  2. 2025UniLat3D: Geometry-Appearance Unified Latents for Single-Stage 3D Generation
  3. 2025ArtiLatent: Realistic Articulated 3D Object Generation via Structured Latents
  4. 2024MeshFormer: High-Quality Mesh Generation with 3D-Guided Reconstruction Model
  5. 2024CraftsMan3D: High-fidelity Mesh Generation with 3D Native Generation and Interactive Geometry Refiner
  6. 2024Part-aware Shape Generation with Latent 3D Diffusion of Neural Voxel Fields
  7. 2022Deep Generative Models on 3D Representations: A Survey
  8. 2024Compress3D: a Compressed Latent Space for 3D Generation from a Single Image