mlsanityData to 14 Oct 2026

FastMap: Revisiting Structure from Motion through First-Order Optimization

arXiv 2505.04612 ↗First posted 7 May 2025cs.CVcited by 8

Not on the map. Why?

Abstract (arXiv)

We propose FastMap, a new global structure from motion method focused on speed and simplicity. Previous methods like COLMAP and GLOMAP are able to estimate high-precision camera poses, but suffer from poor scalability when the number of matched keypoint pairs becomes large, mainly due to the time-consuming process of second-order Gauss-Newton optimization. Instead, we design our method solely based on first-order optimizers. To obtain maximal speedup, we identify and eliminate two key performance bottlenecks: computational complexity and the kernel implementation of each optimization step. Through extensive experiments, we show that FastMap is up to 10 times faster than COLMAP and GLOMAP with GPU acceleration and achieves comparable pose accuracy.

Key facts. Cited by 8 papers in the map (data through 14 Oct 2026). Builds on 43 of its 73 references that are in the map. First posted on arXiv 7 May 2025.

Builds on

43 of 73 references.

References · 43

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  2. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  3. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,226
  4. 2022Instant Neural Graphics Primitives with a Multiresolution Hash Encodingcited by 3,723
  5. 2021Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fieldscited by 2,032
  6. 2023DUSt3R: Geometric 3D Vision Made Easycited by 1,567
  7. 2022Objaverse: A Universe of Annotated 3D Objectscited by 1,540
  8. 2025VGGT: Visual Geometry Grounded Transformercited by 1,539
  9. 2019Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfercited by 1,495
  10. 2018On the Continuity of Rotation Representations in Neural Networkscited by 1,214
  11. 2024Grounding Image Matching in 3D with MASt3Rcited by 983
  12. 2023MVDream: Multi-view Diffusion for 3D Generationcited by 938
31 more
  1. 2024Structured 3D Latents for Scalable and Versatile 3D Generationcited by 770
  2. 2021DROID-SLAM: Deep Visual SLAM for Monocular, Stereo, and RGB-D Camerascited by 687
  3. 2023Zip-NeRF: Anti-Aliased Grid-Based Neural Radiance Fieldscited by 671
  4. 2023ZoeDepth: Zero-shot Transfer by Combining Relative and Metric Depthcited by 657
  5. 2023CoTracker: It is Better to Track Togethercited by 522
  6. 2020TartanAir: A Dataset to Push the Limits of Visual SLAMcited by 506
  7. 2021Mega-NeRF: Scalable Construction of Large-Scale NeRFs for Virtual Fly-Throughscited by 398
  8. 2022Kubric: A scalable dataset generatorcited by 381
  9. 2023Instant3D: Fast Text-to-3D with Sparse-View Generation and Large Reconstruction Modelcited by 378
  10. 2025Fast3R: Towards 3D Reconstruction of 1000+ Images in One Forward Passcited by 367
  11. 2023PointOdyssey: A Large-Scale Synthetic Dataset for Long-Term Point Trackingcited by 276
  12. 2023Visual Geometry Grounded Deep Structure From Motioncited by 227
  13. 2024Global Structure-from-Motion Revisitedcited by 218
  14. 2021Pixel-Perfect Structure-from-Motion with Featuremetric Refinementcited by 200
  15. 2023Spring: A High-Resolution High-Detail Dataset and Benchmark for Scene Flow, Optical Flow and Stereocited by 172
  16. 2024MASt3R-SfM: a Fully-Integrated Solution for Unconstrained Structure-from-Motioncited by 162
  17. 2021Capturing, Reconstructing, and Simulating: the UrbanScene3D Datasetcited by 118
  18. 2021NeRF for Outdoor Scene Relightingcited by 114
  19. 2022LaMAR: Benchmarking Localization and Mapping for Augmented Realitycited by 100
  20. 2023VR-NeRF: High-Fidelity Virtualized Walkable Spacescited by 86
  21. 2024Scene Coordinate Reconstruction: Posing of Image Collections via Incremental Learning of a Relocalizercited by 73
  22. 2024FlowMap: High-Quality Camera Poses, Intrinsics, and Depth via Gradient Descentcited by 61
  23. 2023CamP: Camera Preconditioning for Neural Radiance Fieldscited by 55
  24. 2014Robust Camera Location Estimation by Convex Programmingcited by 41
  25. 2023All for One, and One for All: UrbanSyn Dataset, the third Musketeer of Synthetic Driving Scenescited by 41
  26. 2025Matrix3D: Large Photogrammetry Model All-in-Onecited by 40
  27. 2019Baseline Desensitizing In Translation Averagingcited by 27
  28. 2024NerfBaselines: Consistent and Reproducible Evaluation of Novel View Synthesis Methodscited by 26
  29. 2017A Minimal Solution for Two-view Focal-length Estimation using Two Affine Correspondencescited by 14
  30. 2020On the Distribution of Minima in Intrinsic-Metric Rotation Averagingcited by 9
  31. 2023Robust Self-calibration of Focal Lengths from the Fundamental Matrixcited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 6.

8 papers.

  1. 2025Do You Know Where Your Camera Is? View-Invariant Policy Learning with Camera Conditioningcited by 14
  2. 2025VGGT-X: When VGGT Meets Dense Novel View Synthesiscited by 8
  3. 2026Marginalized Bundle Adjustment: Multi-View Camera Pose from Monocular Depth Estimatescited by 3
  4. 2026Efficient 3D Content Reconstruction and Generationcited by 0
  5. 2026PixVOD: Pixel-Distributed Direct Visual Odometry and Depth Estimationcited by 0
  6. 2026EPO: Boosting 3D Foundation Models with Edge-based Pose Optimizationcited by 0
  7. 2026SalientGS: Unified SfM-to-3DGS with Importance-Guided MCMC Gaussian Allocationcited by 0
  8. 2026Triangulation-Free Bundle Adjustment with Graduated Non-Convexity for Camera Pose Refinement from Coarse Priorscited by 0