mlsanityData to 7 Oct 2026

GeodesicNVS: Probability Density Geodesic Flow Matching for Novel View Synthesis

arXiv 2603.01010 ↗First posted 1 Mar 2026cs.CVcited by 2

On the map: Diffusion models → Flow matching, time series and science

Abstract (arXiv)

Recent advances in generative modeling have substantially enhanced novel view synthesis, yet maintaining consistency across viewpoints remains challenging. Diffusion-based models rely on stochastic noise-to-data transitions, which obscure deterministic structures and yield inconsistent view predictions. We advocate a Data-to-Data Flow Matching framework that learns deterministic transformations between paired views, enhancing view-consistent synthesis through explicit data coupling. Building on this, we propose Probability Density Geodesic Flow Matching (PDG-FM), which aligns interpolation trajectories with density-based geodesics of a data manifold. To enable tractable geodesic estimation, we employ a teacher-student framework that distills density-based geodesic interpolants into an efficient ambient-space predictor. Empirically, our method surpasses diffusion-based baselines on Objaverse and GSO30 datasets, demonstrating improved structural coherence and smoother transitions across views. These results highlight the advantages of incorporating data-dependent geometric regularization into deterministic flow matching for consistent novel view generation.

Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Placed in Flow matching, time series and science, one of 14 directions in Diffusion models. Builds on 46 of its 58 references that are in the map. First posted on arXiv 1 Mar 2026.

Builds on

46 of 58 references.

Closest ancestors · 3

Same direction, newest first

  1. 2024Metric Flow Matching for Smooth Interpolations on the Data Manifold23 May
  2. 2022Flow Matching for Generative Modeling6 Oct
  3. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow7 Sep

Other references · 43

  1. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  3. 2020Denoising Diffusion Implicit Modelscited by 9,655
  4. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 5,986
  5. 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesiscited by 4,290
  6. 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,067
  7. 2022DreamFusion: Text-to-3D using 2D Diffusioncited by 3,006
  8. 2020RAFT: Recurrent All-Pairs Field Transforms for Optical Flowcited by 2,385
  9. 2022An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversioncited by 2,231
  10. 2022Objaverse: A Universe of Annotated 3D Objectscited by 1,522
  11. 2023Zero-1-to-3: Zero-shot One Image to 3D Objectcited by 1,483
  12. 2021Cascaded Diffusion Models for High Fidelity Image Generationcited by 1,155
31 more
  1. 2022Building Normalizing Flows with Stochastic Interpolantscited by 964
  2. 2023SyncDreamer: Generating Multiview-consistent Images from a Single-view Imagecited by 638
  3. 2022Google Scanned Objects: A High-Quality Dataset of 3D Scanned Household Itemscited by 601
  4. 2021Diffusion Schr"odinger Bridge with Applications to Score-Based Generative Modelingcited by 594
  5. 2023Make-It-3D: High-Fidelity 3D Creation from A Single Image with Diffusion Priorcited by 341
  6. 2021Light Field Networks: Neural Scene Representations with Single-Evaluation Renderingcited by 330
  7. 2023RealFusion: 360deg Reconstruction of Any Object from a Single Imagecited by 305
  8. 2023Denoising Diffusion Bridge Modelscited by 171
  9. 2017Latent Space Oddity: on the Curvature of Deep Generative Modelscited by 169
  10. 2024EscherNet: A Generative Model for Scalable View Synthesiscited by 94
  11. 2023Free3D: Consistent Novel View Synthesis without 3D Representationcited by 72
  12. 2025Hume: Introducing System-2 Thinking in Visual-Language-Action Modelcited by 56
  13. 2023Boosting Latent Diffusion with Flow Matchingcited by 44
  14. 2024Neural Flow Diffusion Models: Learnable Forward Process for Improved Diffusion Modellingcited by 43
  15. 2025FlowR: Flowing from Sparse to Dense 3D Reconstructionscited by 28
  16. 2024Unsupervised Hyperspectral Image Super-Resolution via Self-Supervised Modality Decouplingcited by 26
  17. 2021Pulling back information geometrycited by 23
  18. 2022On Investigating the Conservative Property of Score-Based Generative Modelscited by 16
  19. 2024NVComposer: Boosting Generative Novel View Synthesis with Multiple Sparse and Unposed Imagescited by 12
  20. 2024Learning Distances from Data with Normalizing Flows and Score Matchingcited by 10
  21. 2025Probability Density Geodesics in Image Diffusion Latent Spacecited by 9
  22. 2026Omni2Sound: Towards Unified Video-Text-to-Audio Generationcited by 8
  23. 2025In-Context Brush: Zero-shot Customized Subject Insertion with Context-Aware Latent Space Manipulationcited by 7
  24. 2025LADB: Latent Aligned Diffusion Bridges for Semi-Supervised Domain Translationcited by 7
  25. 2026TAG-MoE: Task-Aware Gating for Unified Generative Mixture-of-Expertscited by 6
  26. 2025Latent Swap Joint Diffusion for 2D Long-Form Latent Generationcited by 5
  27. 2026Unified Thinker: A General Reasoning Modular Core for Image Generationcited by 5
  28. 2026Driving with DINO: Vision Foundation Features as a Unified Bridge for Sim-to-Real Generation in Autonomous Drivingcited by 5
  29. 2023Light Field Diffusion for Single-View Novel View Synthesiscited by 3
  30. 2025MeSS: City Mesh-Guided Outdoor Scene Generation with Cross-View Consistent Diffusioncited by 3
  31. 2026DAV-GSWT: Diffusion-Active-View Sampling for Data-Efficient Gaussian Splatting Wang Tilescited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026SpatialReward: Verifiable Spatial Reward Modeling for Fine-Grained Spatial Consistency in Text-to-Image Generationcited by 4
  2. 2026CNMBI: Determining the Number of Clusters Using Center Pairwise Matching and Boundary Filteringcited by 1

Similar, not cited

Close in wording, no citation either way

  1. 2026Probabilistic Geodesic Flow Matching on Location-Scale Families
  2. 2026GeoVerse: World-Consistent Novel View Synthesis in Geometric Latent Space
  3. 2026Flow Matching is Adaptive to Manifold Structures
  4. 2026Repurposing Geometric Foundation Models for Multi-view Diffusion
  5. 2024Flow map matching with stochastic interpolants: A mathematical framework for consistency models
  6. 2026Flow Matching in Feature Space for Stochastic World Modeling
  7. 2026GeoNVS: Geometry Grounded Video Diffusion for Novel View Synthesis
  8. 2026VGGT-Diff: Visual Geometry Meets Diffusion for Sparse-View Novel View Synthesis