mlsanityData to 14 Oct 2026

GenFusion: Closing the Loop between Reconstruction and Generation via Videos

arXiv 2503.21219 ↗First posted 27 Mar 2025cs.CV cs.AIcited by 41

Not on the map. Why?

Abstract (arXiv)

Recently, 3D reconstruction and generation have demonstrated impressive novel view synthesis results, achieving high fidelity and efficiency. However, a notable conditioning gap can be observed between these two fields, e.g., scalable 3D scene reconstruction often requires densely captured views, whereas 3D generation typically relies on a single or no input view, which significantly limits their applications. We found that the source of this phenomenon lies in the misalignment between 3D constraints and generative priors. To address this problem, we propose a reconstruction-driven video diffusion model that learns to condition video frames on artifact-prone RGB-D renderings. Moreover, we propose a cyclical fusion pipeline that iteratively adds restoration frames from the generative model to the training set, enabling progressive expansion and addressing the viewpoint saturation limitations seen in previous reconstruction and generation pipelines. Our evaluation, including view synthesis from sparse view and masked input, validates the effectiveness of our approach. More details at https://genfusion.sibowu.com.

Key facts. Cited by 41 papers in the map (data through 14 Oct 2026). Builds on 59 of its 67 references that are in the map. First posted on arXiv 27 Mar 2025.

Builds on

59 of 67 references.

References · 59

  1. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  2. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,226
  3. 2022DreamFusion: Text-to-3D using 2D Diffusioncited by 3,025
  4. 2023Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasetscited by 2,581
  5. 2021Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fieldscited by 2,032
  6. 2020pixelNeRF: Neural Radiance Fields from One or Few Imagescited by 1,584
  7. 2019Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfercited by 1,495
  8. 20242D Gaussian Splatting for Geometrically Accurate Radiance Fieldscited by 1,076
  9. 2021PlenOctrees for Real-time Rendering of Neural Radiance Fieldscited by 905
  10. 2020NeRF++: Analyzing and Improving Neural Radiance Fieldscited by 861
  11. 2021MVSNeRF: Fast Generalizable Radiance Field Reconstruction from Multi-View Stereocited by 740
  12. 2021Depth-supervised NeRF: Fewer Views and Faster Training for Freecited by 735
38 more
  1. 2023pixelSplat: 3D Gaussian Splats from Image Pairs for Scalable Generalizable 3D Reconstructioncited by 684
  2. 2023Zip-NeRF: Anti-Aliased Grid-Based Neural Radiance Fieldscited by 671
  3. 2021BARF: Bundle-Adjusting Neural Radiance Fieldscited by 642
  4. 2021RegNeRF: Regularizing Neural Radiance Fields for View Synthesis from Sparse Inputscited by 615
  5. 2024MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-View Imagescited by 566
  6. 2023DL3DV-10K: A Large-Scale Scene Dataset for Deep Learning-based 3D Visioncited by 537
  7. 2023DynamiCrafter: Animating Open-domain Images with Video Diffusion Priorscited by 471
  8. 2022MonoSDF: Exploring Monocular Geometric Cues for Neural Implicit Surface Reconstructioncited by 470
  9. 2021Baking Neural Radiance Fields for Real-Time View Synthesiscited by 468
  10. 2023Neuralangelo: High-Fidelity Neural Surface Reconstructioncited by 467
  11. 2024CAT3D: Create Anything in 3D with Multi-View Diffusion Modelscited by 432
  12. 2024ViewCrafter: Taming Video Diffusion Models for High-fidelity Novel View Synthesiscited by 403
  13. 2023ReconFusion: 3D Reconstruction with Diffusion Priorscited by 318
  14. 2023FreeNeRF: Improving Few-shot Neural Rendering with Free Frequency Regularizationcited by 311
  15. 2021Dense Depth Priors for Neural Radiance Fields from Sparse Input Viewscited by 302
  16. 2022Novel View Synthesis with Diffusion Modelscited by 300
  17. 2024DepthSplat: Connecting Gaussian Splatting and Depthcited by 286
  18. 2023FSGS: Real-Time Few-shot View Synthesis using Gaussian Splattingcited by 279
  19. 2024DNGaussian: Optimizing Sparse-View 3D Gaussian Radiance Fields with Global-Local Depth Normalizationcited by 256
  20. 2022Geo-Neus: Geometry-Consistent Neural Implicit Surfaces Learning for Multi-view Reconstructioncited by 212
  21. 2024DepthCrafter: Generating Consistent Long Depth Sequences for Open-world Videoscited by 204
  22. 2021GeoNeRF: Generalizing NeRF with Geometry Priorscited by 196
  23. 2023Single-Stage Diffusion NeRF: A Unified Approach to 3D Generation and Reconstructioncited by 180
  24. 2023WonderJourney: Going from Anywhere to Everywherecited by 158
  25. 2021Efficient Neural Radiance Fields for Interactive Free-viewpoint Videocited by 150
  26. 2021Improving neural implicit surfaces geometry with patch warpingcited by 137
  27. 2023DiffusioNeRF: Regularizing Neural Radiance Fields with Denoising Diffusion Modelscited by 127
  28. 2024Flash3D: Feed-Forward Generalisable 3D Scene Reconstruction from a Single Imagecited by 116
  29. 2024Long-LRM: Long-sequence Large Reconstruction Model for Wide-coverage Gaussian Splatscited by 114
  30. 2023ZeroNVS: Zero-Shot 360-Degree View Synthesis from a Single Imagecited by 111
  31. 20243DGS-Enhancer: Enhancing Unbounded 3D Gaussian Splatting with View-consistent 2D Diffusion Priorscited by 106
  32. 2024FreGS: 3D Gaussian Splatting with Progressive Frequency Regularizationcited by 105
  33. 2022InfiniteNature-Zero: Learning Perpetual View Generation of Natural Scenes from Single Imagescited by 82
  34. 2023Nerfbusters: Removing Ghostly Artifacts from Casually Captured NeRFscited by 76
  35. 2023MuRF: Multi-Baseline Radiance Fieldscited by 71
  36. 2023I2V-Adapter: A General Image-to-Video Adapter for Diffusion Modelscited by 70
  37. 20233D-aware Image Generation using 2D Diffusion Modelscited by 69
  38. 2024GaussianObject: High-Quality 3D Object Reconstruction from Four Views with Gaussian Splattingcited by 59

50 most cited of 59 All 59 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 14; 2026: 27.

41 papers.

Sort:

By area:

  1. 2025WorldMem: Long-term Consistent World Simulation with Memorycited by 168
  2. 2025GSFixer: Improving 3D Gaussian Splatting with Reference-Guided Video Diffusion Priorscited by 38
  3. 2025Advances in Feed-Forward 3D Reconstruction and View Synthesis: A Surveycited by 27
  4. 2025GSFix3D: Diffusion-Guided Repair of Novel Views in Gaussian Splattingcited by 18
  5. 2026World-R1: Reinforcing 3D Constraints for Text-to-Video Generationcited by 11
  6. 2025G4Splat: Geometry-Guided Gaussian Splatting with Generative Priorcited by 10
  7. 2026GaussFusion: Improving 3D Reconstruction in the Wild with A Geometry-Informed Video Generatorcited by 7
  8. 2026A Mechanistic View on Video Generation as World Models: State and Dynamicscited by 6
  9. 2026ArtifactWorld: Scaling 3D Gaussian Splatting Artifact Restoration via Video Generation Modelscited by 5
  10. 2025ProSplat: Improved Feed-Forward 3D Gaussian Splatting for Wide-Baseline Sparse Viewscited by 4
  11. 2025Sparse-View 3D Reconstruction: Recent Advances and Open Challengescited by 4
  12. 2025GaMO: Geometry-aware Multi-view Diffusion Outpainting for Sparse-View 3D Reconstructioncited by 4
29 more
  1. 2026GIFSplat: Generative Prior-Guided Iterative Feed-Forward 3D Gaussian Splatting from Sparse Viewscited by 4
  2. 2026Pantheon360: Taming Digital Twin Generation via 3D-Aware 360deg Video Diffusioncited by 4
  3. 2025FixingGS: Enhancing 3D Gaussian Splatting via Training-Free Score Distillationcited by 3
  4. 2025WaveletGaussian: Wavelet-domain Diffusion for Sparse-view 3D Gaussian Object Reconstructioncited by 3
  5. 2025Self-Evolving 3D Scene Generation from a Single Imagecited by 3
  6. 2025Breaking the Vicious Cycle: Coherent 3D Gaussian Splatting from Sparse and Motion-Blurred Viewscited by 3
  7. 2026GeoNVS: Geometry Grounded Video Diffusion for Novel View Synthesiscited by 3
  8. 2025Droplet3D: Commonsense Priors from Videos Facilitate 3D Generationcited by 1
  9. 2026ConfCtrl: Enabling Precise Camera Control in Video Diffusion via Confidence-Aware Interpolationcited by 1
  10. 2026GSCompleter: A Distillation-Free Plugin for Metric-Aware 3D Gaussian Splatting Completion in Secondscited by 1
  11. 2026HAD: Hallucination-Aware Diffusion Priors for 3D Reconstructioncited by 1
  12. 2026RealityBridge: Bridging Editable 3D Gaussian Splatting Driving Simulations and Real-World Videoscited by 1
  13. 2026Latent Riemannian Flow Matching for Geometry-Grounded 3D Foundation Modelscited by 1
  14. 2026R2S-EGO: Dual-Proxy Refinement for Sparse-Capture Real-to-Simcited by 1
  15. 2025SCORP: Scene-Consistent Object Refinement via Proxy Generation and Tuningcited by 0
  16. 2026Pi-GS: Sparse-View Gaussian Splatting with Dense π3 Initializationcited by 0
  17. 2026CrowdGaussian: Reconstructing High-Fidelity 3D Gaussians for Human Crowd from a Single Imagecited by 0
  18. 2026Can These Views Be One Scene? Evaluating Multiview 3D Consistency when 3D Foundation Models Hallucinatecited by 0
  19. 2026Lighting-Consistent Object Transfer Across Radiance Fieldscited by 0
  20. 2026CORGI: Consistency-Aware 3D Dog Reconstruction from a Single Image in the Wildcited by 0
  21. 2026FillGS: Filling Observation Gaps in 4D Gaussian Splatting via Viewpoint-Time Selection and Generative Refinementcited by 0
  22. 2026TRACE-GS: On-Policy Trajectory Distillation with Privileged Geometric Conditioning for Sparse-View 3DGS Restorationcited by 0
  23. 2026GenRec: Knowing Where to Reconstruct and Where to Generatecited by 0
  24. 2026DualDiff3D: Dual Structure-Appearance Diffusion Priors for Reliability-Enhanced 3D Gaussian Splattingcited by 0
  25. 2026RoGe: Novel View Synthesis via End-to-End Implicit Reconstruction and Generationcited by 0
  26. 2026Rethinking 3D Noise: Learning 3D-Aware Video Priors via Optimization-Free Morphological Perturbationscited by 0
  27. 2026Bi-FlowGS: Bridging Generative View Completion and Gaussian Geometry through Bidirectional Flow Co-Refinementcited by 0
  28. 2026D3GS: Depth, DINO, and RGB Diffusion Co-Guided 3D Gaussian Splatting for Sparse-View Reconstructioncited by 0
  29. 2026ClearGS: Reliability-Aware Gaussian Splatting from Handheld Videoscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Gen3R: 3D Scene Generation Meets Feed-Forward Reconstruction
  2. 2022SparseFusion: Distilling View-conditioned Diffusion for 3D Reconstruction
  3. 2026ArtiFixer: Enhancing and Extending 3D Reconstruction with Auto-Regressive Diffusion Models
  4. 2025VideoScene: Distilling Video Diffusion Model to Generate 3D Scenes in One Step
  5. 2026FixAnything: 3D-Consistent Rendering Refinement via Video Generative Priors
  6. 2026Closing the Loop: Training-Free Revisit Consistency for Autoregressive Generative Rendering
  7. 2024VFusion3D: Learning Scalable 3D Generative Models from Video Diffusion Models
  8. 2026Efficient Camera-Controlled Video Generation of Static Scenes via Sparse Diffusion and 3D Rendering