mlsanityData to 14 Oct 2026

Splatt3R: Zero-shot Gaussian Splatting from Uncalibrated Image Pairs

arXiv 2408.13912 ↗First posted 25 Aug 2024cs.CV cs.LGcited by 178

Not on the map. Why?

Abstract (arXiv)

In this paper, we introduce Splatt3R, a pose-free, feed-forward method for in-the-wild 3D reconstruction and novel view synthesis from stereo pairs. Given uncalibrated natural images, Splatt3R can predict 3D Gaussian Splats without requiring any camera parameters or depth information. For generalizability, we build Splatt3R upon a ``foundation'' 3D geometry reconstruction method, MASt3R, by extending it to deal with both 3D structure and appearance. Specifically, unlike the original MASt3R which reconstructs only 3D point clouds, we predict the additional Gaussian attributes required to construct a Gaussian primitive for each point. Hence, unlike other novel view synthesis methods, Splatt3R is first trained by optimizing the 3D point cloud's geometry loss, and then a novel view synthesis objective. By doing this, we avoid the local minima present in training 3D Gaussian Splats from stereo views. We also propose a novel loss masking strategy that we empirically find is critical for strong performance on extrapolated viewpoints. We train Splatt3R on the ScanNet++ dataset and demonstrate excellent generalisation to uncalibrated, in-the-wild images. Splatt3R can reconstruct scenes at 4FPS at 512 x 512 resolution, and the resultant splats can be rendered in real-time.

Key facts. Cited by 178 papers in the map (data through 14 Oct 2026). Builds on 46 of its 68 references that are in the map. First posted on arXiv 25 Aug 2024.

Builds on

46 of 68 references.

References · 46

  1. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  2. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,226
  3. 2022Instant Neural Graphics Primitives with a Multiresolution Hash Encodingcited by 3,723
  4. 2021Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fieldscited by 2,032
  5. 2021Vision Transformers for Dense Predictioncited by 1,803
  6. 2020pixelNeRF: Neural Radiance Fields from One or Few Imagescited by 1,584
  7. 2023DUSt3R: Geometric 3D Vision Made Easycited by 1,567
  8. 2017SuperPoint: Self-Supervised Interest Point Detection and Descriptioncited by 1,320
  9. 2024Grounding Image Matching in 3D with MASt3Rcited by 983
  10. 2019Scene Representation Networks: Continuous 3D-Structure-Aware Neural Scene Representationscited by 864
  11. 2016Unsupervised Monocular Depth Estimation with Left-Right Consistencycited by 820
  12. 2023ScanNet++: A High-Fidelity Dataset of 3D Indoor Scenescited by 764
34 more
  1. 2021MVSNeRF: Fast Generalizable Radiance Field Reconstruction from Multi-View Stereocited by 740
  2. 2021IBRNet: Learning Multi-View Image-Based Renderingcited by 703
  3. 2023pixelSplat: 3D Gaussian Splats from Image Pairs for Scalable Generalizable 3D Reconstructioncited by 684
  4. 2021Common Objects in 3D: Large-Scale Learning and Evaluation of Real-life 3D Category Reconstructioncited by 652
  5. 2021BARF: Bundle-Adjusting Neural Radiance Fieldscited by 642
  6. 2024MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-View Imagescited by 566
  7. 2018Pyramid Stereo Matching Networkcited by 489
  8. 2016Unsupervised CNN for Single View Depth Estimation: Geometry to the Rescuecited by 368
  9. 2023Splatter Image: Ultra-Fast Single-View 3D Reconstructioncited by 364
  10. 2018GeoNet: Unsupervised Learning of Dense Depth, Optical Flow and Camera Posecited by 342
  11. 2021Light Field Networks: Neural Scene Representations with Single-Evaluation Renderingcited by 332
  12. 2019GA-Net: Guided Aggregation Net for End-to-end Stereo Matchingcited by 244
  13. 2022NoPe-NeRF: Optimising Neural Radiance Field with No Pose Priorcited by 235
  14. 2021Self-Calibrating Neural Radiance Fieldscited by 204
  15. 2021GeoNeRF: Generalizing NeRF with Geometry Priorscited by 196
  16. 2022SPARF: Neural Radiance Fields from Sparse and Noisy Posescited by 195
  17. 2021Neural Rays for Occlusion-aware Image-based Renderingcited by 192
  18. 2021Stereo Radiance Fields (SRF): Learning View Synthesis for Sparse Views of Novel Scenescited by 190
  19. 2022SparseNeuS: Fast Generalizable Neural Surface Reconstruction from Sparse Viewscited by 177
  20. 2023GPS-Gaussian: Generalizable Pixel-wise 3D Gaussian Splatting for Real-time Human Novel View Synthesiscited by 170
  21. 2018Unsupervised Learning of Monocular Depth Estimation and Visual Odometry with Deep Feature Reconstructioncited by 159
  22. 2022Generalizable Patch-Based Neural Renderingcited by 122
  23. 2024latentSplat: Autoencoding Variational Gaussians for Fast Generalizable 3D Reconstructioncited by 118
  24. 2024Flash3D: Feed-Forward Generalisable 3D Scene Reconstruction from a Single Imagecited by 116
  25. 2023Learning to Render Novel Views from Wide-Baseline Stereo Pairscited by 104
  26. 2014Computing the Stereo Matching Cost with a Convolutional Neural Networkcited by 74
  27. 2023LEAP: Liberate Sparse-view 3D Modeling from Camera Posescited by 65
  28. 2023DBARF: Deep Bundle-Adjusting Generalizable Neural Radiance Fieldscited by 60
  29. 2021Deep Two-View Structure-from-Motion Revisitedcited by 39
  30. 2023FlowCam: Training Generalizable 3D Radiance Fields without Camera Poses via Pixel-Aligned Scene Flowcited by 39
  31. 2023PoRF: Pose Residual Field for Accurate Neural Surface Reconstructioncited by 19
  32. 2023ColNeRF: Collaboration for Generalizable Sparse Input Neural Radiance Fieldcited by 11
  33. 2024Generalizable Novel-View Synthesis using a Stereo Cameracited by 5
  34. 2020Large Scale Photometric Bundle Adjustmentcited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 19; 2025: 79; 2026: 80.

178 papers.

Sort:

By area:

  1. 2025Depth Anything 3: Recovering the Visual Space from Any Viewscited by 692
  2. 2024No Pose, No Problem: Surprisingly Simple 3D Gaussian Splats from Sparse Unposed Imagescited by 248
  3. 2025AnySplat: Feed-forward 3D Gaussian Splatting from Unconstrained Viewscited by 243
  4. 2025FLARE: Feed-forward Geometry, Appearance and Camera Estimation from Uncalibrated Sparse Viewscited by 180
  5. 2025Streaming 4D Visual Geometry Transformercited by 161
  6. 2024MV-DUSt3R+: Single-Stage Scene Reconstruction from Sparse Views In 2 Secondscited by 141
  7. 2025VGGT-SLAM: Dense RGB SLAM Optimized on the SL(4) Manifoldcited by 136
  8. 2024MVSplat360: Feed-Forward 360 Scene Synthesis from Sparse Viewscited by 116
  9. 2024SLAM3R: Real-Time Dense Scene Reconstruction from Monocular RGB Videoscited by 102
  10. 2025Geometry Forcing: Marrying Video Diffusion and 3D Representation for Consistent World Modelingcited by 96
  11. 2024Reloc3r: Large-Scale Training of Relative Camera Pose Regression for Generalizable, Fast, and Accurate Visual Localizationcited by 82
  12. 2025Pow3R: Empowering Unconstrained 3D Reconstruction with Camera and Scene Priorscited by 72
38 more
  1. 2025Geo4D: Leveraging Video Generators for Geometric 4D Scene Reconstructioncited by 67
  2. 2025WorldMirror: Universal 3D World Reconstruction with Any-Prior Promptingcited by 67
  3. 2024PF3plat: Pose-Free Feed-Forward 3D Gaussian Splattingcited by 62
  4. 2024SelfSplat: Pose-Free and 3D Prior-Free Generalizable 3D Gaussian Splattingcited by 42
  5. 2024FreeSplatter: Pose-free Gaussian Splatting for Sparse-view 3D Reconstructioncited by 41
  6. 2025No Pose at All: Self-Supervised Pose-Free 3D Gaussian Splatting from Sparse Viewscited by 40
  7. 2025GWM: Towards Scalable Gaussian World Models for Robotic Manipulationcited by 39
  8. 2025YoNoSplat: You Only Need One Model for Feedforward 3D Gaussian Splattingcited by 37
  9. 2026LingBot-Map: Geometric Context Transformer for Streaming 3D Reconstructioncited by 37
  10. 2024PhysFlow: Unleashing the Potential of Multi-modal Foundation Models and Video Diffusion for 4D Dynamic Physical Scene Simulationcited by 36
  11. 2025MoVieS: Motion-Aware 4D Dynamic View Synthesis in One Secondcited by 35
  12. 2025ReconViaGen: Towards Accurate Multi-view 3D Object Reconstruction via Generationcited by 33
  13. 2025WonderTurbo: Generating Interactive 3D World in 0.72 Secondscited by 32
  14. 2024PreF3R: Pose-Free Feed-Forward 3D Gaussian Splatting from Variable-length Image Sequencecited by 28
  15. 20244Real-Video: Learning Generalizable Photo-Realistic 4D Video Diffusioncited by 27
  16. 2024Feat2GS: Probing Visual Foundation Models with Gaussian Splattingcited by 27
  17. 2025Advances in Feed-Forward 3D Reconstruction and View Synthesis: A Surveycited by 27
  18. 2025Uni3R: Unified 3D Reconstruction and Semantic Understanding via Generalizable Gaussian Splatting from Unposed Multi-View Imagescited by 26
  19. 2024PanSplat: 4K Panorama Synthesis with Feed-Forward Gaussian Splattingcited by 24
  20. 2025Zero-Shot Monocular Scene Flow Estimation in the Wildcited by 22
  21. 2025EVolSplat: Efficient Volume-based Gaussian Splatting for Urban View Synthesiscited by 21
  22. 2025C3G: Learning Compact 3D Representations with 2K Gaussianscited by 21
  23. 2025UniLat3D: Geometry-Appearance Unified Latents for Single-Stage 3D Generationcited by 20
  24. 2025A Survey on 3D Gaussian Splatting Applications: Segmentation, Editing, and Generationcited by 19
  25. 2025LiteVGGT: Boosting Vanilla VGGT via Geometry-aware Cached Token Mergingcited by 17
  26. 2025SpatialCrafter: Unleashing the Imagination of Video Diffusion Models for Scene Reconstruction from Limited Observationscited by 16
  27. 2025SIU3R: Simultaneous Scene Understanding and 3D Reconstruction Beyond Feature Alignmentcited by 16
  28. 2024GSemSplat: Generalizable Semantic 3D Gaussian Splatting from Uncalibrated Image Pairscited by 15
  29. 2025VicaSplat: A Single Run is All You Need for 3D Gaussian Splatting and Camera Estimation from Unposed Video Framescited by 15
  30. 2025SteerX: Creating Any Camera-Free 3D and 4D Scenes with Geometric Steeringcited by 15
  31. 2025FreeSplat++: Generalizable 3D Gaussian Splatting for Efficient Indoor Scene Reconstructioncited by 15
  32. 2025HAMSt3R: Human-Aware Multi-view Stereo 3D Reconstructioncited by 15
  33. 2025UniForward: Unified 3D Scene and Semantic Field Reconstruction via Feed-Forward Gaussian Splatting from Only Sparse-View Imagescited by 14
  34. 20254Real-Video-V2: Fused View-Time Attention and Feedforward Reconstruction for 4D Scene Generationcited by 14
  35. 2026TokenSplat: Token-aligned 3D Gaussian Splatting for Feed-forward Pose-free Reconstructioncited by 14
  36. 2024Latent-Reframe: Enabling Camera Control for Video Diffusion Model without Trainingcited by 13
  37. 2025SpatialSplat: Efficient Semantic 3D from Sparse Unposed Imagescited by 13
  38. 2025RegGS: Unposed Sparse Views Gaussian Splatting with 3DGS Registrationcited by 13

50 most cited of 178 All 178 (JSON)