mlsanityData to 7 Oct 2026

TrackNeRF: Bundle Adjusting NeRF from Sparse and Noisy Views via Feature Tracks

arXiv 2408.10739 ↗First posted 20 Aug 2024cs.CVcited by 12

Not on the map. Why?

Abstract (arXiv)

Neural radiance fields (NeRFs) generally require many images with accurate poses for accurate novel view synthesis, which does not reflect realistic setups where views can be sparse and poses can be noisy. Previous solutions for learning NeRFs with sparse views and noisy poses only consider local geometry consistency with pairs of views. Closely following textitbundle adjustment in Structure-from-Motion (SfM), we introduce TrackNeRF for more globally consistent geometry reconstruction and more accurate pose optimization. TrackNeRF introduces textitfeature tracks, ie connected pixel trajectories across textitall visible views that correspond to the textitsame 3D points. By enforcing reprojection consistency among feature tracks, TrackNeRF encourages holistic 3D consistency explicitly. Through extensive experiments, TrackNeRF sets a new benchmark in noisy and sparse view reconstruction. In particular, TrackNeRF shows significant improvements over the state-of-the-art BARF and SPARF by ~8 and ~1 in terms of PSNR on DTU under various sparse and noisy view setups. The code is available at hrefhttps://tracknerf.github.io/.

Key facts. Cited by 12 papers in the map (data through 7 Oct 2026). Builds on 81 of its 103 references that are in the map. First posted on arXiv 20 Aug 2024.

Builds on

81 of 103 references.

References · 81

  1. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
  2. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,141
  3. 2022Instant Neural Graphics Primitives with a Multiresolution Hash Encodingcited by 3,698
  4. 2022DreamFusion: Text-to-3D using 2D Diffusioncited by 3,006
  5. 2020RAFT: Recurrent All-Pairs Field Transforms for Optical Flowcited by 2,385
  6. 2021Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fieldscited by 2,013
  7. 2021Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fieldscited by 1,957
  8. 2021Plenoxels: Radiance Fields without Neural Networkscited by 1,607
  9. 2020pixelNeRF: Neural Radiance Fields from One or Few Imagescited by 1,574
  10. 2023DUSt3R: Geometric 3D Vision Made Easycited by 1,526
  11. 2023Zero-1-to-3: Zero-shot One Image to 3D Objectcited by 1,483
  12. 2017SuperPoint: Self-Supervised Interest Point Detection and Descriptioncited by 1,313
38 more
  1. 2022Magic3D: High-Resolution Text-to-3D Content Creationcited by 1,285
  2. 2018On the Continuity of Rotation Representations in Neural Networkscited by 1,186
  3. 2016Unsupervised Monocular Depth Estimation with Left-Right Consistencycited by 819
  4. 2018MVSNet: Depth Inference for Unstructured Multi-view Stereocited by 780
  5. 2021Depth-supervised NeRF: Fewer Views and Faster Training for Freecited by 732
  6. 2023Fantasia3D: Disentangling Geometry and Appearance for High-quality Text-to-3D Content Creationcited by 684
  7. 2021DROID-SLAM: Deep Visual SLAM for Monocular, Stereo, and RGB-D Camerascited by 673
  8. 2019Local Light Field Fusion: Practical View Synthesis with Prescriptive Sampling Guidelinescited by 672
  9. 2023pixelSplat: 3D Gaussian Splats from Image Pairs for Scalable Generalizable 3D Reconstructioncited by 668
  10. 2021BARF: Bundle-Adjusting Neural Radiance Fieldscited by 641
  11. 2021RegNeRF: Regularizing Neural Radiance Fields for View Synthesis from Sparse Inputscited by 614
  12. 2021NICE-SLAM: Neural Implicit Scalable Encoding for SLAMcited by 586
  13. 2021Ref-NeRF: Structured View-Dependent Appearance for Neural Radiance Fieldscited by 569
  14. 2024MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-View Imagescited by 553
  15. 2023CoTracker: It is Better to Track Togethercited by 516
  16. 2021Putting NeRF on a Diet: Semantically Consistent Few-Shot View Synthesiscited by 464
  17. 2023Magic123: One Image to High-Quality 3D Object Generation Using Both 2D and 3D Diffusion Priorscited by 395
  18. 2023Splatter Image: Ultra-Fast Single-View 3D Reconstructioncited by 359
  19. 2018GeoNet: Unsupervised Learning of Dense Depth, Optical Flow and Camera Posecited by 341
  20. 2023ReconFusion: 3D Reconstruction with Diffusion Priorscited by 317
  21. 2023FreeNeRF: Improving Few-shot Neural Rendering with Free Frequency Regularizationcited by 310
  22. 2023RealFusion: 360deg Reconstruction of Any Object from a Single Imagecited by 305
  23. 2023TAPIR: Tracking Any Point with per-frame Initialization and temporal Refinementcited by 281
  24. 2023FSGS: Real-Time Few-shot View Synthesis using Gaussian Splattingcited by 278
  25. 2018DeepMVS: Learning Multi-view Stereopsiscited by 277
  26. 2024DNGaussian: Optimizing Sparse-View 3D Gaussian Radiance Fields with Global-Local Depth Normalizationcited by 256
  27. 2019Recurrent MVSNet for High-resolution Multi-view Stereo Depth Inferencecited by 252
  28. 2023SparseNeRF: Distilling Depth Ranking for Few-shot Novel View Synthesiscited by 247
  29. 2023Visual Geometry Grounded Deep Structure From Motioncited by 225
  30. 2021Self-Calibrating Neural Radiance Fieldscited by 204
  31. 2021Pixel-Perfect Structure-from-Motion with Featuremetric Refinementcited by 199
  32. 2022SPARF: Neural Radiance Fields from Sparse and Noisy Posescited by 195
  33. 2022NeRF-SLAM: Real-Time Dense Monocular SLAM with Neural Radiance Fieldscited by 188
  34. 2021InfoNeRF: Ray Entropy Minimization for Few-Shot Neural Volume Renderingcited by 180
  35. 2023PF-LRM: Pose-Free Large Reconstruction Model for Joint Pose and Shape Predictioncited by 165
  36. 2023SparseGS: Sparse View Synthesis using 3D Gaussian Splattingcited by 134
  37. 2020RAFT-3D: Scene Flow using Rigid-Motion Embeddingscited by 121
  38. 2020Fast-MVSNet: Sparse-to-Dense Multi-View Stereo With Learned Propagation and Gauss-Newton Refinementcited by 105

50 most cited of 81 All 81 (JSON)

Built on by

Cited by year

Citing papers per year: 2024: 4; 2025: 8.

12 papers.

  1. 2024Wonderland: Navigating 3D Scenes from a Single Imagecited by 98
  2. 20243D Convex Splatting: Radiance Field Rendering with 3D Smooth Convexescited by 50
  3. 2025Can Video Diffusion Model Reconstruct 4D Geometry?cited by 10
  4. 2024Generic Objects as Pose Probes for Few-shot View Synthesiscited by 9
  5. 2025VideoLifter: Lifting Videos to 3D with Fast Hierarchical Stereo Alignmentcited by 8
  6. 2024FusionRF: High-Fidelity Satellite Neural Radiance Fields from Multispectral and Panchromatic Acquisitionscited by 5
  7. 2025TrackGS: Optimizing COLMAP-Free 3D Gaussian Splatting with Global Track Constraintscited by 4
  8. 2025UKBOB: One Billion MRI Labeled Masks for Generalizable 3D Medical Image Segmentationcited by 4
  9. 2025UFV-Splatter: Pose-Free Feed-Forward 3D Gaussian Splatting Adapted to Unfavorable Viewscited by 3
  10. 2025Weakly-Supervised Learning of Dense Functional Correspondencescited by 2
  11. 2025Learning Fine-Grained Geometry for Sparse-View Splatting via Cascade Depth Losscited by 1
  12. 2025NoPe-NeRF++: Local-to-Global Optimization of NeRF with No Pose Priorcited by 1