mlsanityData to 14 Oct 2026

Bilevel Online Adaptation for Out-of-Domain Human Mesh Reconstruction

arXiv 2103.16449 ↗First posted 30 Mar 2021cs.CVcited by 39

Not on the map. Why?

Abstract (arXiv)

This paper considers a new problem of adapting a pre-trained model of human mesh reconstruction to out-of-domain streaming videos. However, most previous methods based on the parametric SMPL model citeloper2015smpl underperform in new domains with unexpected, domain-specific attributes, such as camera parameters, lengths of bones, backgrounds, and occlusions. Our general idea is to dynamically fine-tune the source model on test video streams with additional temporal constraints, such that it can mitigate the domain gaps without over-fitting the 2D information of individual test frames. A subsequent challenge is how to avoid conflicts between the 2D and temporal constraints. We propose to tackle this problem using a new training algorithm named Bilevel Online Adaptation (BOA), which divides the optimization process of overall multi-objective into two steps of weight probe and weight update in a training iteration. We demonstrate that BOA leads to state-of-the-art results on two human mesh reconstruction benchmarks.

Key facts. Cited by 39 papers in the map (data through 14 Oct 2026). Builds on 45 of its 65 references that are in the map. First posted on arXiv 30 Mar 2021.

Builds on

45 of 65 references.

References · 45

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  3. 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 7,005
  4. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,897
  5. 2019Expressive Body Capture: 3D Hands, Face, and Body from a Single Imagecited by 1,612
  6. 2018Group Normalizationcited by 1,524
  7. 2017Deeper, Broader and Artier Domain Generalizationcited by 986
  8. 2017End-to-end Recovery of Human Shape and Posecited by 885
  9. 2017Learning to Generalize: Meta-Learning for Domain Generalizationcited by 763
  10. 2016Keep it SMPL: Automatic Estimation of 3D Human Pose and Shape from a Single Imagecited by 695
  11. 2019Learning to Reconstruct 3D Human Pose and Shape via Model-fitting in the Loopcited by 627
  12. 2013Domain Generalization via Invariant Feature Representationcited by 623
33 more
  1. 2019VIBE: Video Inference for Human Body Pose and Shape Estimationcited by 620
  2. 2016Monocular 3D Human Pose Estimation In The Wild Using Improved CNN Supervisioncited by 494
  3. 2018Generalizing to Unseen Domains via Adversarial Data Augmentationcited by 451
  4. 2018Learning 3D Human Dynamics from Videocited by 316
  5. 2020I2L-MeshNet: Image-to-Lixel Prediction Network for Accurate 3D Human Pose and Mesh Estimation from a Single RGB Imagecited by 274
  6. 2019Convolutional Mesh Regression for Single-Image Human Shape Reconstructioncited by 270
  7. 2018Generalizing Across Domains via Cross-Gradient Trainingcited by 268
  8. 2015Domain Generalization for Object Recognition with Multi-task Autoencoderscited by 253
  9. 2020Pose2Mesh: Graph Convolutional Network for 3D Human Pose and Mesh Recovery from a 2D Human Posecited by 248
  10. 2019Episodic Training for Domain Generalizationcited by 245
  11. 2018Learning to Estimate 3D Human Pose and Shape from a Single Color Imagecited by 226
  12. 2017Unite the People: Closing the Loop Between 3D and 2D Human Representationscited by 205
  13. 2018Neural Body Fitting: Unifying Deep Learning and Model-Based Human Pose and Shape Estimationcited by 185
  14. 2019On the Convergence Theory of Gradient-Based Model-Agnostic Meta-Learning Algorithmscited by 132
  15. 2020Exemplar Fine-Tuning for 3D Human Model Fitting Towards In-the-Wild 3D Human Pose Estimationcited by 113
  16. 2019Exploiting temporal context for 3D human pose estimation in the wildcited by 110
  17. 2017Online Adaptation of Convolutional Neural Networks for Video Object Segmentationcited by 109
  18. 2019DenseRaC: Joint 3D Pose and Shape Estimation by Dense Render-and-Comparecited by 100
  19. 2017Self-supervised Learning of Motion Capturecited by 97
  20. 2019Structured Prediction Helps 3D Human Motion Modellingcited by 96
  21. 2017Towards Accurate Markerless Human Shape and Pose Estimation over Timecited by 92
  22. 2019Human Mesh Recovery from Monocular Images via a Skeleton-disentangled Representationcited by 87
  23. 2020Human Body Model Fitting by Learned Gradient Descentcited by 81
  24. 2018Real-time self-adaptive deep stereocited by 80
  25. 2019Sim2real transfer learning for 3D human pose estimation: motion to the rescuecited by 76
  26. 2013Streaming Variational Bayescited by 48
  27. 2019Learning to Adapt for Stereocited by 42
  28. 2020Predicting Camera Viewpoint Improves Cross-dataset Generalization for 3D Human Pose Estimationcited by 36
  29. 2020Self-Supervised Deep Visual Odometry with Online Adaptationcited by 35
  30. 2020Inference Stage Optimization for Cross-scenario 3D Human Pose Estimationcited by 35
  31. 2018Meta-Tracker: Fast and Robust Online Adaptation for Visual Object Trackerscited by 34
  32. 2020Chained Representation Cycling: Learning to Estimate 3D Human Pose and Shape by Cycling Between Representationscited by 12
  33. 2019Towards Robust RGB-D Human Mesh Recoverycited by 3

Built on by

Cited by year

Citing papers per year: 2021: 4; 2022: 10; 2023: 11; 2024: 10; 2025: 4.

39 papers.

Sort:

By area:

  1. 2023A Comprehensive Survey on Test-Time Adaptation under Distribution Shiftscited by 359
  2. 2022CLIFF: Carrying Location Information in Full Frames into Human Pose and Shape Estimationcited by 238
  3. 2022PyMAF-X: Towards Well-aligned Full-body Model Regression from Monocular Imagescited by 153
  4. 2022Recovering 3D Human Mesh from Monocular Images: A Surveycited by 103
  5. 20233D Human Pose Estimation via Intuitive Physicscited by 103
  6. 2021Playing for 3D Human Recoverycited by 64
  7. 2023EgoHumans: An Egocentric 3D Multi-Human Benchmarkcited by 60
  8. 2022Occluded Human Mesh Recoverycited by 54
  9. 2022GIPSO: Geometrically Informed Propagation for Online Adaptation in 3D LiDAR Segmentationcited by 35
  10. 2021AdaptPose: Cross-Dataset Adaptation for 3D Human Pose Estimation by Learnable Motion Generationcited by 32
  11. 2022MeshMAE: Masked Autoencoders for 3D Mesh Data Analysiscited by 31
  12. 2024Harmony4D: A Video Dataset for In-The-Wild Close Human Interactionscited by 31
27 more
  1. 2024PhysPT: Physics-aware Pretrained Transformer for Estimating Human Dynamics from Monocular Videoscited by 28
  2. 2022Uncertainty-Aware Adaptation for Self-Supervised 3D Human Pose Estimationcited by 25
  3. 2021Out-of-Domain Human Mesh Reconstruction via Dynamic Bilevel Online Adaptationcited by 24
  4. 2023Global Adaptation meets Local Generalization: Unsupervised Domain Adaptation for 3D Human Pose Estimationcited by 19
  5. 2023ZeroAvatar: Zero-shot 3D Avatar Generation from a Single Imagecited by 18
  6. 2023Cyclic Test-Time Adaptation on Monocular Video for 3D Human Mesh Reconstructioncited by 18
  7. 2022Domain Adaptive 3D Pose Augmentation for In-the-wild Human Mesh Recoverycited by 12
  8. 2023JOTR: 3D Joint Contrastive Learning with Transformers for Occluded Human Mesh Recoverycited by 12
  9. 2023PoSynDA: Multi-Hypothesis Pose Synthesis Domain Adaptation for Robust 3D Human Pose Estimationcited by 12
  10. 2021Lifting 2D Human Pose to 3D with Domain Adapted 3D Body Conceptcited by 4
  11. 2022Skeleton2Humanoid: Animating Simulated Characters for Physically-plausible Motion In-betweeningcited by 4
  12. 2024MoCap-to-Visual Domain Adaptation for Efficient Human Mesh Estimation from 2D Keypointscited by 3
  13. 2022NeMo: 3D Neural Motion Fields from Multiple Video Instances of the Same Actioncited by 2
  14. 2023STRIDE: Single-video based Temporally Continuous Occlusion-Robust 3D Pose Estimationcited by 2
  15. 2024Uncertainty-Aware Testing-Time Optimization for 3D Human Pose Estimationcited by 2
  16. 2024Personalized 3D Human Pose and Shape Refinementcited by 2
  17. 2024Incorporating Test-Time Optimization into Training with Dual Networks for Human Mesh Recoverycited by 1
  18. 2024ESCAPE: Energy-based Selective Adaptive Correction for Out-of-distribution 3D Human Pose Estimationcited by 1
  19. 2024Motion Diffusion-Guided 3D Global HMR from a Dynamic Cameracited by 1
  20. 2025Semantics-aware Test-time Adaptation for 3D Human Pose Estimationcited by 1
  21. 2025Online Test-time Adaptation for 3D Human Pose Estimation: A Practical Perspective with Estimated 2D Posescited by 1
  22. 2025Unsupervised Cross-Domain 3D Human Pose Estimation via Pseudo-Label-Guided Global Transformscited by 1
  23. 2023Mirror-Aware Neural Humanscited by 0
  24. 20233DHR-Co: A Collaborative Test-time Refinement Framework for In-the-Wild 3D Human-Body Reconstruction Taskcited by 0
  25. 2024RemoCap: Disentangled Representation Learning for Motion Capturecited by 0
  26. 2024Estimating Human Poses Across Datasets: A Unified Skeleton and Multi-Teacher Distillation Approachcited by 0
  27. 2025Robust Long-term Test-Time Adaptation for 3D Human Pose Estimation through Motion Discretizationcited by 0