mlsanityData to 14 Oct 2026

WorldMirror: Universal 3D World Reconstruction with Any-Prior Prompting

arXiv 2510.10726 ↗First posted 12 Oct 2025cs.CVcited by 67

Not on the map. Why?

Abstract (arXiv)

We present WorldMirror, a unified feed-forward model for comprehensive 3D geometric prediction tasks. Unlike existing methods constrained to image-only inputs or customized for a specific task, our framework flexibly integrates diverse geometric priors, including camera poses, intrinsics, and depth maps, while simultaneously generating multiple 3D representations: dense point clouds, multi-view depth maps, camera parameters, surface normals, and 3D Gaussians. Remarkably, prior injection yields universal gains across all tasks, suggesting that input flexibility and multi-task prediction are mutually reinforcing. WorldMirror achieves state-of-the-art performance across diverse benchmarks from camera, point map, depth, and surface normal estimation to novel view synthesis, while maintaining the efficiency of feed-forward inference.

Key facts. Cited by 67 papers in the map (data through 14 Oct 2026). Builds on 65 of its 83 references that are in the map. First posted on arXiv 12 Oct 2025.

Builds on

65 of 83 references.

References · 65

  1. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,226
  2. 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,117
  3. 2021Vision Transformers for Dense Predictioncited by 1,803
  4. 2020pixelNeRF: Neural Radiance Fields from One or Few Imagescited by 1,584
  5. 2023DUSt3R: Geometric 3D Vision Made Easycited by 1,567
  6. 2025VGGT: Visual Geometry Grounded Transformercited by 1,539
  7. 2017Matterport3D: Learning from RGB-D Data in Indoor Environmentscited by 1,376
  8. 2024Depth Anything: Unleashing the Power of Large-Scale Unlabeled Datacited by 1,360
  9. 2017Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semanticscited by 1,358
  10. 2019Habitat: A Platform for Embodied AI Researchcited by 1,191
  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. 2023pixelSplat: 3D Gaussian Splats from Image Pairs for Scalable Generalizable 3D Reconstructioncited by 684
  3. 2023Mip-Splatting: Alias-free 3D Gaussian Splattingcited by 667
  4. 2021Common Objects in 3D: Large-Scale Learning and Evaluation of Real-life 3D Category Reconstructioncited by 652
  5. 2018Taskonomy: Disentangling Task Transfer Learningcited by 634
  6. 2020Hypersim: A Photorealistic Synthetic Dataset for Holistic Indoor Scene Understandingcited by 604
  7. 2024MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-View Imagescited by 566
  8. 2025Continuous 3D Perception Model with Persistent Statecited by 548
  9. 2023DL3DV-10K: A Large-Scale Scene Dataset for Deep Learning-based 3D Visioncited by 537
  10. 2023CoTracker: It is Better to Track Togethercited by 522
  11. 2020TartanAir: A Dataset to Push the Limits of Visual SLAMcited by 506
  12. 2019BlendedMVS: A Large-scale Dataset for Generalized Multi-view Stereo Networkscited by 490
  13. 2024CameraCtrl: Enabling Camera Control for Text-to-Video Generationcited by 460
  14. 2021ARKitScenes: A Diverse Real-World Dataset For 3D Indoor Scene Understanding Using Mobile RGB-D Datacited by 454
  15. 2024MonST3R: A Simple Approach for Estimating Geometry in the Presence of Motioncited by 442
  16. 2024UniDepth: Universal Monocular Metric Depth Estimationcited by 382
  17. 2025Fast3R: Towards 3D Reconstruction of 1000+ Images in One Forward Passcited by 367
  18. 2021Omnidata: A Scalable Pipeline for Making Multi-Task Mid-Level Vision Datasets from 3D Scanscited by 336
  19. 2025MapAnything: Universal Feed-Forward Metric 3D Reconstructioncited by 314
  20. 2021Neural RGB-D Surface Reconstructioncited by 313
  21. 2024GeoWizard: Unleashing the Diffusion Priors for 3D Geometry Estimation from a Single Imagecited by 286
  22. 2024DepthSplat: Connecting Gaussian Splatting and Depthcited by 286
  23. 2024gsplat: An Open-Source Library for Gaussian Splattingcited by 285
  24. 2018DeepMVS: Learning Multi-view Stereopsiscited by 277
  25. 2024No Pose, No Problem: Surprisingly Simple 3D Gaussian Splats from Sparse Unposed Imagescited by 248
  26. 2025AnySplat: Feed-forward 3D Gaussian Splatting from Unconstrained Viewscited by 243
  27. 2023MatrixCity: A Large-scale City Dataset for City-scale Neural Rendering and Beyondcited by 185
  28. 2025FLARE: Feed-forward Geometry, Appearance and Camera Estimation from Uncalibrated Sparse Viewscited by 180
  29. 2024Splatt3R: Zero-shot Gaussian Splatting from Uncalibrated Image Pairscited by 178
  30. 2023Aria Digital Twin: A New Benchmark Dataset for Egocentric 3D Machine Perceptioncited by 167
  31. 2023PF-LRM: Pose-Free Large Reconstruction Model for Joint Pose and Shape Predictioncited by 165
  32. 2024StableNormal: Reducing Diffusion Variance for Stable and Sharp Normalcited by 154
  33. 2023PoseDiffusion: Solving Pose Estimation via Diffusion-aided Bundle Adjustmentcited by 147
  34. 2018Evaluation of CNN-based Single-Image Depth Estimation Methodscited by 138
  35. 2024Rethinking Inductive Biases for Surface Normal Estimationcited by 124
  36. 2021SMD-Nets: Stereo Mixture Density Networkscited by 120
  37. 20223D Common Corruptions and Data Augmentationcited by 119
  38. 2021Estimating and Exploiting the Aleatoric Uncertainty in Surface Normal Estimationcited by 116

50 most cited of 65 All 65 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 11; 2026: 56.

67 papers.

Sort:

By area:

  1. 2025WorldPlay: Towards Long-Term Geometric Consistency for Real-Time Interactive World Modelingcited by 170
  2. 2025Hunyuan-GameCraft-2: Instruction-following Interactive Game World Modelcited by 61
  3. 2026NeoVerse: Enhancing 4D World Model with in-the-wild Monocular Videoscited by 46
  4. 2026HY-World 2.0: A Multi-Modal World Model for Reconstructing, Generating, and Simulating 3D Worldscited by 44
  5. 2026Latent-WAM: Latent World Action Modeling for End-to-End Autonomous Drivingcited by 33
  6. 2026VerseCrafter: Dynamic Realistic Video World Model with 4D Geometric Controlcited by 28
  7. 2026WorldCompass: Reinforcement Learning for Long-Horizon World Modelscited by 24
  8. 2026WorldStereo: Bridging Camera-Guided Video Generation and Scene Reconstruction via 3D Geometric Memoriescited by 18
  9. 2025EcoSplat: Efficiency-controllable Feed-forward 3D Gaussian Splatting from Multi-view Imagescited by 10
  10. 2026UCM: Unified Modeling of Camera Control and Memory with Time-aware Positional Encoding Warping for World Modelscited by 9
  11. 2025Unified Semantic Transformer for 3D Scene Understandingcited by 8
  12. 2026AnchorSplat: Feed-Forward 3D Gaussian Splatting with 3D Geometric Priorscited by 8
38 more
  1. 2026WorldClaw: Agentic 3D Open-World Generation at Scalecited by 8
  2. 2025The World is Your Canvas: Painting Promptable Events with Reference Images, Trajectories, and Textcited by 7
  3. 2025MVInverse: Feed-forward Multi-view Inverse Rendering in Secondscited by 6
  4. 2026Feed-Forward 3D Scene Modeling: A Problem-Driven Perspectivecited by 6
  5. 2026From Rays to Projections: Better Inputs for Feed-Forward View Synthesiscited by 5
  6. 2025Selfi: Self Improving Reconstruction Engine via 3D Geometric Feature Alignmentcited by 4
  7. 2026SoMA: A Real-to-Sim Neural Simulator for Robotic Soft-body Manipulationcited by 4
  8. 2026Large Depth Completion Model from Sparse Observationscited by 4
  9. 2026OmniX: Any-view and Any-time 4D Reconstruction via Feed-forward Trajectory Fieldscited by 4
  10. 2025GeoWorld: Providing Full-frame Geometry Features to Facilitate 3D Scene Generationcited by 3
  11. 2025Self-Evolving 3D Scene Generation from a Single Imagecited by 3
  12. 2026Efficient Camera-Controlled Video Generation of Static Scenes via Sparse Diffusion and 3D Renderingcited by 3
  13. 2026Research on World Models Is Not Merely Injecting World Knowledge into Specific Taskscited by 3
  14. 2026HD-VGGT: High-Resolution Visual Geometry Transformercited by 3
  15. 2026OpenWorldLib: A Unified Codebase and Definition of Advanced World Modelscited by 3
  16. 2026Holo-World: Unified Camera, Object and Weather Control for Video World Modelcited by 3
  17. 2026Real-Time Human Frontal View Synthesis from a Single Imagecited by 2
  18. 2026SplatWeaver: Learning to Allocate Gaussian Primitives for Generalizable Novel View Synthesiscited by 2
  19. 2026OnlineWM: Causality-Aware Active Online Learning for Effective World Modelingcited by 2
  20. 2025CustomX: Unified Character, Action, and Scene Customization in Video World Modelscited by 1
  21. 2026AirSplat: Alignment and Rating for Robust Feed-Forward 3D Gaussian Splattingcited by 1
  22. 2026Learn2Splat: Extending the Horizon of Learned 3DGS Optimizationcited by 1
  23. 2026IVGT: Implicit Visual Geometry Transformer for Neural Scene Representationcited by 1
  24. 2026SceneAligner: 3D-Grounded Floorplan Localization in the Wildcited by 1
  25. 2026OF3GS: On-the-Fly Feed-Forward 3D Gaussian Splatting from Unposed Imagescited by 1
  26. 2026Argus: Metric Panoramic 3D Reconstruction for Indoor Scenescited by 1
  27. 2026Beyond 2D Matching: A Unified Single-Stage Framework for Geometry-Aware Cross-View Object Geo-Localizationcited by 1
  28. 2026InvSplat: Inverse Feed-Forward Scene Splattingcited by 1
  29. 2026ABot-3DWorld 0: A Universal World Model to Explore Any 3D Spacecited by 1
  30. 2026R2S-EGO: Dual-Proxy Refinement for Sparse-Capture Real-to-Simcited by 1
  31. 2026GeoFF3D: Coordinate-Anchored Feed-Forward Reconstruction for Large-Scale UAV Mappingcited by 1
  32. 2026G-ray: Ray-Level Relative Geometric Position Encoding in Multi-View Vision Transformers under Camera Heterogeneitycited by 1
  33. 2025Quantile Rendering: Efficiently Embedding High-dimensional Feature on 3D Gaussian Splattingcited by 0
  34. 2026MV-SAM: Multi-view Promptable Segmentation using Pointmap Guidancecited by 0
  35. 2026SceneExpander: Text-Guided 3D Scene Expansion via Free-Form View Insertioncited by 0
  36. 2026ConFixGS: Learning to Fix Feedforward 3D Gaussian Splatting with Confidence-Aware Diffusion Priors in Driving Scenescited by 0
  37. 2026Cross-View Splatter: Feed-Forward View Synthesis with Georeferenced Imagescited by 0
  38. 2026UniT: Unified Geometry Learning with Group Autoregressive Transformercited by 0

50 most cited of 67 All 67 (JSON)