mlsanityData to 7 Oct 2026

Learning from Synthetic Humans

arXiv 1701.01370 ↗First posted 5 Jan 2017cs.CVcited by 358

Not on the map. Why?

Abstract (arXiv)

Estimating human pose, shape, and motion from images and videos are fundamental challenges with many applications. Recent advances in 2D human pose estimation use large amounts of manually-labeled training data for learning convolutional neural networks (CNNs). Such data is time consuming to acquire and difficult to extend. Moreover, manual labeling of 3D pose, depth and motion is impractical. In this work we present SURREAL (Synthetic hUmans foR REAL tasks): a new large-scale dataset with synthetically-generated but realistic images of people rendered from 3D sequences of human motion capture data. We generate more than 6 million frames together with ground truth pose, depth maps, and segmentation masks. We show that CNNs trained on our synthetic dataset allow for accurate human depth estimation and human part segmentation in real RGB images. Our results and the new dataset open up new possibilities for advancing person analysis using cheap and large-scale synthetic data.

Key facts. Cited by 358 papers in the map (data through 7 Oct 2026). First posted on arXiv 5 Jan 2017.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 89; 2022: 70; 2023: 77; 2024: 66; 2025: 29; 2026: 27.

358 papers.

Sort:

By area:

  1. 2023Synthetic Data from Diffusion Models Improves ImageNet Classificationcited by 326
  2. 2021HuMoR: 3D Human Motion Model for Robust Pose Estimationcited by 319
  3. 2021A-NeRF: Articulated Neural Radiance Fields for Learning Human Shape, Appearance, and Posecited by 233
  4. 2023BEDLAM: A Synthetic Dataset of Bodies Exhibiting Detailed Lifelike Animated Motioncited by 229
  5. 2021AGORA: Avatars in Geography Optimized for Regression Analysiscited by 203
  6. 2021Fake It Till You Make It: Face analysis in the wild using synthetic data alonecited by 200
  7. 2021Neural Articulated Radiance Fieldcited by 196
  8. 2022Embodied Hands: Modeling and Capturing Hands and Bodies Togethercited by 179
  9. 2023StableRep: Synthetic Images from Text-to-Image Models Make Strong Visual Representation Learnerscited by 177
  10. 2022Fake it till you make it: Learning transferable representations from synthetic ImageNet clonescited by 171
  11. 2022InstantAvatar: Learning Avatars from Monocular Video in 60 Secondscited by 138
  12. 2021SPEC: Seeing People in the Wild with an Estimated Cameracited by 134
38 more
  1. 2021Generative Models as a Data Source for Multiview Representation Learningcited by 104
  2. 2022Recovering 3D Human Mesh from Monocular Images: A Surveycited by 103
  3. 2022Generative Neural Articulated Radiance Fieldscited by 101
  4. 2021Pre-training without Natural Imagescited by 100
  5. 2024Multi-HMR: Multi-Person Whole-Body Human Mesh Recovery in a Single Shotcited by 98
  6. 2023EMDB: The Electromagnetic Database of Global 3D Human Pose and Shape in the Wildcited by 95
  7. 2021PoseAug: A Differentiable Pose Augmentation Framework for 3D Human Pose Estimationcited by 88
  8. 2021NeuroMorph: Unsupervised Shape Interpolation and Correspondence in One Gocited by 81
  9. 2021SCALE: Modeling Clothed Humans with a Surface Codec of Articulated Local Elementscited by 72
  10. 2023Learning Vision from Models Rivals Learning Vision from Datacited by 71
  11. 2023High-fidelity 3D Human Digitization from Single 2K Resolution Imagescited by 70
  12. 2021MOTSynth: How Can Synthetic Data Help Pedestrian Detection and Tracking?cited by 68
  13. 2023Human-Art: A Versatile Human-Centric Dataset Bridging Natural and Artificial Scenescited by 68
  14. 2022Unsupervised Learning of Efficient Geometry-Aware Neural Articulated Representationscited by 65
  15. 2024SynthCLIP: Are We Ready for a Fully Synthetic CLIP Training?cited by 64
  16. 2021Playing for 3D Human Recoverycited by 63
  17. 2024HumanVid: Demystifying Training Data for Camera-controllable Human Image Animationcited by 63
  18. 2023SINC: Spatial Composition of 3D Human Motions for Simultaneous Action Generationcited by 62
  19. 2021Neural Re-Rendering of Humans from a Single Imagecited by 59
  20. 2021S3: Neural Shape, Skeleton, and Skinning Fields for 3D Human Modelingcited by 59
  21. 2022A Survey of Non-Rigid 3D Registrationcited by 59
  22. 2023EgoHumans: An Egocentric 3D Multi-Human Benchmarkcited by 58
  23. 2023Unsupervised Learning of Robust Spectral Shape Matchingcited by 57
  24. 2024Neural Localizer Fields for Continuous 3D Human Pose and Shape Estimationcited by 57
  25. 2022UnrealEgo: A New Dataset for Robust Egocentric 3D Human Motion Capturecited by 56
  26. 2022AvatarCLIP: Zero-Shot Text-Driven Generation and Animation of 3D Avatarscited by 55
  27. 2023SynBody: Synthetic Dataset with Layered Human Models for 3D Human Perception and Modelingcited by 54
  28. 2023Learning Locally Editable Virtual Humanscited by 54
  29. 2021Style and Pose Control for Image Synthesis of Humans from a Single Monocular Viewcited by 53
  30. 2022OPD: Single-view 3D Openable Part Detectioncited by 52
  31. 2021Shape registration in the time of transformerscited by 51
  32. 2022Learning 3D Human Pose Estimation from Dozens of Datasets using a Geometry-Aware Autoencoder to Bridge Between Skeleton Formatscited by 50
  33. 2021Recent Advances in Monocular 2D and 3D Human Pose Estimation: A Deep Learning Perspectivecited by 49
  34. 2022Accurate 3D Body Shape Regression using Metric and Semantic Attributescited by 49
  35. 2021Hierarchical Kinematic Probability Distributions for 3D Human Shape and Pose Estimation from Images in the Wildcited by 48
  36. 2021HumanGAN: A Generative Model of Humans Imagescited by 47
  37. 2022Being Comes from Not-being: Open-vocabulary Text-to-Motion Generation with Wordless Trainingcited by 45
  38. 20233D Human Mesh Estimation from Virtual Markerscited by 44

50 most cited of 358 All 358 (JSON)