mlsanityData to 14 Oct 2026

MapAnything: Universal Feed-Forward Metric 3D Reconstruction

arXiv 2509.13414 ↗First posted 16 Sep 2025cs.CV cs.AI cs.LGcited by 314

Not on the map. Why?

Cited by 26 in World models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

We introduce MapAnything, a unified transformer-based feed-forward model that ingests one or more images along with optional geometric inputs such as camera intrinsics, poses, depth, or partial reconstructions, and then directly regresses the metric 3D scene geometry and cameras. MapAnything leverages a factored representation of multi-view scene geometry, i.e., a collection of depth maps, local ray maps, camera poses, and a metric scale factor that effectively upgrades local reconstructions into a globally consistent metric frame. Standardizing the supervision and training across diverse datasets, along with flexible input augmentation, enables MapAnything to address a broad range of 3D vision tasks in a single feed-forward pass, including uncalibrated structure-from-motion, calibrated multi-view stereo, monocular depth estimation, camera localization, depth completion, and more. We provide extensive experimental analyses and model ablations demonstrating that MapAnything outperforms or matches specialist feed-forward models while offering more efficient joint training behavior, thus paving the way toward a universal 3D reconstruction backbone.

Key facts. Cited by 314 papers in the map (data through 14 Oct 2026). Builds on 67 of its 89 references that are in the map. First posted on arXiv 16 Sep 2025.

Builds on

67 of 89 references.

References · 67

  1. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
  2. 2023Adding Conditional Control to Text-to-Image Diffusion Modelscited by 5,503
  3. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,703
  4. 2021Vision Transformers for Dense Predictioncited by 1,803
  5. 2023DUSt3R: Geometric 3D Vision Made Easycited by 1,567
  6. 2025VGGT: Visual Geometry Grounded Transformercited by 1,539
  7. 2019Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfercited by 1,495
  8. 2019SuperGlue: Learning Feature Matching with Graph Neural Networkscited by 1,443
  9. 2024Depth Anything V2cited by 1,422
  10. 2023T2I-Adapter: Learning Adapters to Dig out More Controllable Ability for Text-to-Image Diffusion Modelscited by 1,351
  11. 2024Grounding Image Matching in 3D with MASt3Rcited by 983
  12. 2023ScanNet++: A High-Fidelity Dataset of 3D Indoor Scenescited by 764
38 more
  1. 2018MegaDepth: Learning Single-View Depth Prediction from Internet Photoscited by 737
  2. 2018Taskonomy: Disentangling Task Transfer Learningcited by 634
  3. 2025Continuous 3D Perception Model with Persistent Statecited by 548
  4. 2023DL3DV-10K: A Large-Scale Scene Dataset for Deep Learning-based 3D Visioncited by 537
  5. 2020TartanAir: A Dataset to Push the Limits of Visual SLAMcited by 506
  6. 2019BlendedMVS: A Large-scale Dataset for Generalized Multi-view Stereo Networkscited by 490
  7. 2024Depth Pro: Sharp Monocular Metric Depth in Less Than a Secondcited by 404
  8. 2024Metric3Dv2: A Versatile Monocular Geometric Foundation Model for Zero-shot Metric Depth and Surface Normal Estimationcited by 395
  9. 2023SplaTAM: Splat, Track & Map 3D Gaussians for Dense RGB-D SLAMcited by 388
  10. 2025Fast3R: Towards 3D Reconstruction of 1000+ Images in One Forward Passcited by 367
  11. 2024MoGe: Unlocking Accurate Monocular Geometry Estimation for Open-Domain Images with Optimal Training Supervisioncited by 299
  12. 2022MultiMAE: Multi-modal Multi-task Masked Autoencoderscited by 291
  13. 2018DeepMVS: Learning Multi-view Stereopsiscited by 277
  14. 20243D Reconstruction with Spatial Memorycited by 277
  15. 2025MoGe-2: Accurate Monocular Geometry with Metric Scale and Sharp Detailscited by 261
  16. 2017A General and Adaptive Robust Loss Functioncited by 236
  17. 2023Visual Geometry Grounded Deep Structure From Motioncited by 227
  18. 2024Global Structure-from-Motion Revisitedcited by 218
  19. 2024MASt3R-SLAM: Real-Time Dense SLAM with 3D Reconstruction Priorscited by 205
  20. 2016DeMoN: Depth and Motion Network for Learning Monocular Stereocited by 201
  21. 2024LVSM: A Large View Synthesis Model with Minimal 3D Inductive Biascited by 189
  22. 2018DeepV2D: Video to Depth with Differentiable Structure from Motioncited by 185
  23. 2025Stable Virtual Camera: Generative View Synthesis with Diffusion Modelscited by 182
  24. 2025UniDepthV2: Universal Monocular Metric Depth Estimation Made Simplercited by 173
  25. 2023Spring: A High-Resolution High-Detail Dataset and Benchmark for Scene Flow, Optical Flow and Stereocited by 172
  26. 2023AM-RADIO: Agglomerative Vision Foundation Model -- Reduce All Domains Into Onecited by 169
  27. 2022CroCo v2: Improved Cross-view Completion Pre-training for Stereo Matching and Optical Flowcited by 166
  28. 2023PF-LRM: Pose-Free Large Reconstruction Model for Joint Pose and Shape Predictioncited by 165
  29. 2024MASt3R-SfM: a Fully-Integrated Solution for Unconstrained Structure-from-Motioncited by 162
  30. 2023DynamicStereo: Consistent Dynamic Depth from Stereo Videoscited by 148
  31. 2024MV-DUSt3R+: Single-Stage Scene Reconstruction from Sparse Views In 2 Secondscited by 141
  32. 2025MUSt3R: Multi-view Network for Stereo 3D Reconstructioncited by 141
  33. 2024Cameras as Rays: Pose Estimation via Ray Diffusioncited by 131
  34. 2021SMD-Nets: Stereo Mixture Density Networkscited by 120
  35. 2024Generative Camera Dolly: Extreme Monocular Dynamic Novel View Synthesiscited by 117
  36. 2020Fast-MVSNet: Sparse-to-Dense Multi-View Stereo With Learned Propagation and Gauss-Newton Refinementcited by 105
  37. 2023Toward General-Purpose Robots via Foundation Models: A Survey and Meta-Analysiscited by 103
  38. 2024SceneScript: Reconstructing Scenes With An Autoregressive Structured Language Modelcited by 96

50 most cited of 67 All 67 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 49; 2026: 265.

314 papers.

Sort:

By area:

  1. 2025Depth Anything 3: Recovering the Visual Space from Any Viewscited by 692
  2. 2025WorldMirror: Universal 3D World Reconstruction with Any-Prior Promptingcited by 67
  3. 2025Efficiently Reconstructing Dynamic Scenes One D4RT at a Timecited by 54
  4. 2025RoMa v2: Harder Better Faster Denser Feature Matchingcited by 51
  5. 2026NeoVerse: Enhancing 4D World Model with in-the-wild Monocular Videoscited by 46
  6. 2025G2VLM: Geometry Grounded Vision Language Model with Unified 3D Reconstruction and Spatial Reasoningcited by 44
  7. 2026HY-World 2.0: A Multi-Modal World Model for Reconstructing, Generating, and Simulating 3D Worldscited by 44
  8. 2025Spatia: Video Generation with Updatable Spatial Memorycited by 40
  9. 2026LingBot-Map: Geometric Context Transformer for Streaming 3D Reconstructioncited by 37
  10. 2025Any4D: Unified Feed-Forward Metric 4D Reconstructioncited by 35
  11. 2026V-DPM: 4D Video Reconstruction with Dynamic Point Mapscited by 34
  12. 2025Trace Anything: Representing Any Video in 4D via Trajectory Fieldscited by 33
38 more
  1. 2026Gen3R: 3D Scene Generation Meets Feed-Forward Reconstructioncited by 31
  2. 2025What Is The Best 3D Scene Representation for Robotics? From Geometric to Foundation Modelscited by 29
  3. 2026VGGT-SLAM 2.0: Real-time Dense Feed-forward Scene Reconstructioncited by 28
  4. 2026LiveWorld: Simulating Out-of-Sight Dynamics in Generative Video World Modelscited by 28
  5. 2025Epipolar Geometry Improves Video Generation Modelscited by 26
  6. 2026ZipMap: Linear-Time Stateful 3D Reconstruction via Test-Time Trainingcited by 26
  7. 2026VGGT-Omegacited by 26
  8. 2026Think3D: Thinking with Space for Spatial Reasoningcited by 24
  9. 2026VGGRPO: Towards World-Consistent Video Generation with 4D Latent Rewardcited by 24
  10. 2026ShapeR: Robust Conditional 3D Shape Generation from Casual Capturescited by 22
  11. 2025DVGT: Driving Visual Geometry Transformercited by 21
  12. 2025AMB3R: Accurate Feed-forward Metric-scale 3D Reconstruction with Backendcited by 20
  13. 2025RealX3D: A Physically-Degraded 3D Benchmark for Multi-view Visual Restoration and Reconstructioncited by 19
  14. 2026VGG-T3: Offline Feed-Forward 3D Reconstruction at Scalecited by 18
  15. 2026WorldStereo: Bridging Camera-Guided Video Generation and Scene Reconstruction via 3D Geometric Memoriescited by 18
  16. 2026Vista4D: Video Reshooting with 4D Point Cloudscited by 18
  17. 2025FlashVGGT: Efficient and Scalable Visual Geometry Transformers with Compressed Descriptor Attentioncited by 17
  18. 2026Latent Spatial Memory for Video World Modelscited by 17
  19. 2025AVGGT: Rethinking Global Attention for Accelerating VGGTcited by 16
  20. 2026Repurposing Geometric Foundation Models for Multi-view Diffusioncited by 15
  21. 2026TokenSplat: Token-aligned 3D Gaussian Splatting for Feed-forward Pose-free Reconstructioncited by 14
  22. 2026Video Generation Models are General-Purpose Vision Learnerscited by 14
  23. 2025In Pursuit of Pixel Supervision for Visual Pre-trainingcited by 13
  24. 2026Any3D-VLA: Enhancing VLA Robustness via Diverse Point Cloudscited by 13
  25. 2026Track4World: Feedforward World-centric Dense 3D Tracking of All Pixelscited by 13
  26. 2026VGGT-World: Transforming VGGT into an Autoregressive Geometry World Modelcited by 13
  27. 2026DVGT-2: Vision-Geometry-Action Model for Autonomous Driving at Scalecited by 13
  28. 2025Towards Spatial Trace with Reasoning in Vision-Language Models for Roboticscited by 12
  29. 2026AnyView: Synthesizing Any Novel View in Dynamic Scenescited by 12
  30. 2026Egocentric World Model for Photorealistic Hand-Object Interaction Synthesiscited by 12
  31. 2026GlobalSplat: Efficient Feed-Forward 3D Gaussian Splatting via Global Scene Tokenscited by 12
  32. 2026MemoBench: Benchmarking World Modeling in Dynamically Changing Environmentscited by 12
  33. 2025Off The Grid: Detection of Primitives for Feed-Forward 3D Gaussian Splattingcited by 11
  34. 2026Geometric Action Model for Robot Policy Learningcited by 11
  35. 2025Taming Camera-Controlled Video Generation with Verifiable Geometry Rewardcited by 10
  36. 2026MoRe: Motion-aware Feed-forward 4D Reconstruction Transformercited by 9
  37. 2025Unified Semantic Transformer for 3D Scene Understandingcited by 8
  38. 2026MVGGT: Multimodal Visual Geometry Grounded Transformer for Multiview 3D Referring Expression Segmentationcited by 8

50 most cited of 314 All 314 (JSON)