mlsanityData to 7 Oct 2026

HOT-POT: Optimal Transport for Sparse Stereo Matching

arXiv 2601.12423 ↗First posted 18 Jan 2026cs.CV math.OCcited by 0

Not on the map. Why?

Abstract (arXiv)

Stereo vision between images faces a range of challenges, including occlusions, motion, and camera distortions, across applications in autonomous driving, robotics, and face analysis. Due to parameter sensitivity, further complications arise for stereo matching with sparse features, such as facial landmarks. To overcome this ill-posedness and enable unsupervised sparse matching, we consider line constraints of the camera geometry from an optimal transport (OT) viewpoint. Formulating camera-projected points as (half)lines, we propose the use of the classical epipolar distance as well as a 3D ray distance to quantify matching quality. Employing these distances as a cost function of a (partial) OT problem, we arrive at efficiently solvable assignment problems. Moreover, we extend our approach to unsupervised object matching by formulating it as a hierarchical OT problem. The resulting algorithms allow for efficient feature and object matching, as demonstrated in our numerical experiments. Here, we focus on applications in facial analysis, where we aim to match distinct landmarking conventions.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 36 of its 78 references that are in the map. First posted on arXiv 18 Jan 2026.

Builds on

36 of 78 references.

References · 36

  1. 2013Sinkhorn Distances: Lightspeed Computation of Optimal Transportation Distancescited by 2,277
  2. 2021LoFTR: Detector-Free Local Feature Matching with Transformerscited by 974
  3. 2018Computational Optimal Transportcited by 765
  4. 2017How far are we from solving the 2D & 3D Face Alignment problem? (and a dataset of 230,000 3D facial landmarks)cited by 524
  5. 2019Generalized Sliced Wasserstein Distancescited by 205
  6. 2021Fake It Till You Make It: Face analysis in the wild using synthetic data alonecited by 200
  7. 2015The MegaFace Benchmark: 1 Million Faces for Recognition at Scalecited by 199
  8. 2020Geometric Dataset Distances via Optimal Transportcited by 157
  9. 2020FLOT: Scene Flow on Point Clouds Guided by Optimal Transportcited by 129
  10. 2019A Wasserstein-type distance in the space of Gaussian Mixture Modelscited by 75
  11. 2018Facial Landmark Detection: a Literature Surveycited by 63
  12. 2019Hierarchical Optimal Transport for Document Representationcited by 53
24 more
  1. 2022Unbalanced Optimal Transport, from Theory to Numericscited by 45
  2. 2014Multi-marginal optimal transport: theory and applicationscited by 41
  3. 2021Self-Point-Flow: Self-Supervised Scene Flow Estimation from Point Clouds with Optimal Transport and Random Walkcited by 39
  4. 2021Accurate Point Cloud Registration with Robust Optimal Transportcited by 38
  5. 2019On the Complexity of Approximating Multimarginal Optimal Transportcited by 27
  6. 2022Sliced Optimal Partial Transportcited by 19
  7. 2024Wasserstein Gradient Flows for Moreau Envelopes of f-Divergences in Reproducing Kernel Hilbert Spacescited by 15
  8. 2019FAB: A Robust Facial Landmark Detection Framework for Motion-Blurred Videoscited by 13
  9. 2021H-Net: Unsupervised Attention-based Stereo Depth Estimation Leveraging Epipolar Geometrycited by 12
  10. 2021Unbalanced Multi-Marginal Optimal Transportcited by 11
  11. 2016DASC: Robust Dense Descriptor for Multi-modal and Multi-spectral Correspondence Estimationcited by 10
  12. 2024Parallelly Sliced Optimal Transport on Spheres and on the Rotation Groupcited by 10
  13. 2025Lightspeed Geometric Dataset Distance via Sliced Optimal Transportcited by 10
  14. 2025Flowing Datasets with Wasserstein over Wasserstein Gradient Flowscited by 10
  15. 2022Lagrangian Motion Magnification with Double Sparse Optical Flow Decompositioncited by 7
  16. 2023On Partial Optimal Transport: Revising the Infeasibility of Sinkhorn and Efficient Gradient Methodscited by 7
  17. 2024T-FAKE: Synthesizing Thermal Images for Facial Landmarkingcited by 7
  18. 2025Slicing the Gaussian Mixture Wasserstein Distancecited by 7
  19. 2022Accelerating the Sinkhorn algorithm for sparse multi-marginal optimal transport by fast Fourier transformscited by 5
  20. 2025Slicing Wasserstein Over Wasserstein Via Functional Optimal Transportcited by 5
  21. 2021Pix2Point: Learning Outdoor 3D Using Sparse Point Clouds and Optimal Transportcited by 4
  22. 2025Structure-Aware Correspondence Learning for Relative Pose Estimationcited by 3
  23. 2022A framework for stereo vision via optimal transportcited by 1
  24. 2022Comparison of Stereo Matching Algorithms for the Development of Disparity Mapcited by 1

Built on by

Not cited yet.