mlsanityData to 14 Oct 2026

Hand-4DGS: Feed-Forward 3D Gaussian Splatting for 4D Hand Reconstruction from Egocentric Videos

arXiv 2606.19156 ↗First posted 17 Jun 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Dynamic 3D hand reconstruction from egocentric videos is essential for next-generation computing platforms such as AR/VR and AI glasses. Despite its importance, most prior works focus either on multi-view 3D hand reconstruction or on 4D human body reconstruction. Egocentric 4D hand reconstruction remains difficult due to rapid hand and camera motion, hand-object and inter-hand interactions, and inherent ambiguity from single-view observations. To address these challenges, we introduce Hand-4DGS, a feed-forward framework for dynamic 4D hand reconstruction from egocentric videos. Our approach incorporates a mesh-guided representation for structural priors and temporal convolutions to model dynamic motion. We evaluate our framework on H2O and ARCTIC, two egocentric hand-object interaction datasets, and show improvements over baselines. Our model can efficiently adapt to unseen videos from datasets that are not included in training. In addition, image supervision through differentiable Gaussian rasterization provides an additional signal for appearance optimization and pose adjustment during training and test-time optimization, without ground-truth 3D hand pose annotations.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 57 of its 68 references that are in the map. First posted on arXiv 17 Jun 2026.

Builds on

57 of 68 references.

References · 57

  1. 2017Attention Is All You Needcited by 60,361
  2. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,226
  3. 2025SAM 3: Segment Anything with Conceptscited by 906
  4. 2023pixelSplat: 3D Gaussian Splats from Image Pairs for Scalable Generalizable 3D Reconstructioncited by 684
  5. 2024MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-View Imagescited by 566
  6. 2020Rescaling Egocentric Visioncited by 540
  7. 2021Zero-Shot Text-Guided Object Generation with Dream Fieldscited by 513
  8. 2019Learning joint reconstruction of hands and manipulated objectscited by 378
  9. 2023Humans in 4D: Reconstructing and Tracking Humans with Transformerscited by 350
  10. 2022ARCTIC: A Dataset for Dexterous Bimanual Hand-Object Manipulationcited by 312
  11. 2023Reconstructing Hands in 3D with Transformerscited by 306
  12. 2021H2O: Two Hands Manipulating Objects for First Person Interaction Recognitioncited by 254
38 more
  1. 2023GPS-Gaussian: Generalizable Pixel-wise 3D Gaussian Splatting for Real-time Human Novel View Synthesiscited by 170
  2. 2021FrankMocap: A Monocular 3D Whole-Body Pose Estimation System via Regression and Integrationcited by 169
  3. 2023HoloAssist: an Egocentric Human Interaction Dataset for Interactive AI Assistants in the Real Worldcited by 166
  4. 2023HUGS: Human Gaussian Splatscited by 154
  5. 2020Reconstructing Hand-Object Interactions in the Wildcited by 144
  6. 2024L4GM: Large 4D Gaussian Reconstruction Modelcited by 144
  7. 2021Semi-Supervised 3D Hand-Object Poses Estimation with Interactions in Timecited by 141
  8. 2024WiLoR: End-to-end 3D Hand Localization and Reconstruction in-the-wildcited by 124
  9. 2024Flash3D: Feed-Forward Generalisable 3D Scene Reconstruction from a Single Imagecited by 116
  10. 2024TACO: Benchmarking Generalizable Bimanual Tool-ACtion-Object Understandingcited by 88
  11. 2022Interacting Attention Graph for Single Image Two-Hand Reconstructioncited by 87
  12. 2023HOLD: Category-agnostic 3D Reconstruction of Interacting Hands and Objects from Videocited by 78
  13. 2025HaWoR: World-Space Hand Motion Reconstruction from Egocentric Videoscited by 66
  14. 2024IDOL: Instant Photorealistic 3D Human Creation from a Single Imagecited by 55
  15. 2024HumanSplat: Generalizable Single-Image Human Gaussian Splatting with Structure Priorscited by 54
  16. 2024Dyn-HaMR: Recovering 4D Interacting Hand Motion from a Dynamic Cameracited by 51
  17. 20254DGT: Learning a 4D Gaussian Transformer Using Real-World Monocular Videoscited by 49
  18. 2024Generalizable Human Gaussians for Sparse View Synthesiscited by 40
  19. 2022Hand Avatar: Free-Pose Hand Animation and Rendering from Monocular Videocited by 37
  20. 2023ACR: Attention Collaboration-based Regressor for Arbitrary Two-Hand Reconstructioncited by 37
  21. 2023Bringing Inputs to Shared Domains for 3D Interacting Hands Recovery in the Wildcited by 36
  22. 2020SeqHAND:RGB-Sequence-Based 3D Hand Pose and Shape Estimationcited by 34
  23. 2022HARP: Personalized Hand Reconstruction from a Monocular RGB Videocited by 28
  24. 2023HandNeRF: Neural Radiance Fields for Animatable Interacting Handscited by 27
  25. 2023LiveHand: Real-time and Photorealistic Neural Hand Renderingcited by 26
  26. 2023HMP: Hand Motion Priors for Pose and Shape Estimation from Videocited by 23
  27. 2025GUAVA: Generalizable Upper Body 3D Gaussian Avatarcited by 22
  28. 2024Expressive Gaussian Human Avatars from Monocular RGB Videocited by 21
  29. 2025Predicting 4D Hand Trajectory from Monocular Videoscited by 21
  30. 2024OHTA: One-shot Hand Avatar via Data-driven Implicit Priorscited by 18
  31. 2022TempCLR: Reconstructing Hands via Time-Coherent Contrastive Learningcited by 17
  32. 2024Authentic Hand Avatar from a Phone Scan via Universal Hand Modelcited by 13
  33. 2024Good, Cheap, and Fast: Overfitted Image Compression with Wasserstein Distortioncited by 13
  34. 2025RoGSplat: Learning Robust Generalizable Human Gaussian Splatting from Sparse Multi-View Imagescited by 12
  35. 2025LIFe-GoM: Generalizable Human Rendering with Learned Iterative Feedback Over Multi-Resolution Gaussians-on-Meshcited by 11
  36. 2025BIGS: Bimanual Category-agnostic Interaction Reconstruction from Monocular Videos via 3D Gaussian Splattingcited by 10
  37. 2023Decoupled Iterative Refinement Framework for Interacting Hands Reconstruction from a Single RGB Imagecited by 8
  38. 2023Wasserstein Distortion: Unifying Fidelity and Realismcited by 8

50 most cited of 57 All 57 (JSON)

Built on by

Not cited yet.