mlsanityData to 14 Oct 2026

X as Supervision: Contending with Depth Ambiguity in Unsupervised Monocular 3D Pose Estimation

arXiv 2411.13026 ↗First posted 20 Nov 2024cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Recent unsupervised methods for monocular 3D pose estimation have endeavored to reduce dependence on limited annotated 3D data, but most are solely formulated in 2D space, overlooking the inherent depth ambiguity issue. Due to the information loss in 3D-to-2D projection, multiple potential depths may exist, yet only some of them are plausible in human structure. To tackle depth ambiguity, we propose a novel unsupervised framework featuring a multi-hypothesis detector and multiple tailored pretext tasks. The detector extracts multiple hypotheses from a heatmap within a local window, effectively managing the multi-solution problem. Furthermore, the pretext tasks harness 3D human priors from the SMPL model to regularize the solution space of pose estimation, aligning it with the empirical distribution of 3D human structures. This regularization is partially achieved through a GCN-based discriminator within the discriminative learning, and is further complemented with synthetic images through rendering, ensuring plausible estimations. Consequently, our approach demonstrates state-of-the-art unsupervised 3D pose estimation performance on various human datasets. Further evaluations on data scale-up and one animal dataset highlight its generalization capabilities. Code will be available at https://github.com/Charrrrrlie/X-as-Supervision.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 40 of its 53 references that are in the map. First posted on arXiv 20 Nov 2024.

Builds on

40 of 53 references.

References · 40

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  3. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  4. 2017Inductive Representation Learning on Large Graphscited by 6,320
  5. 2018Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognitioncited by 1,209
  6. 2016Least Squares Generative Adversarial Networkscited by 1,001
  7. 2016NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysiscited by 686
  8. 2016Monocular 3D Human Pose Estimation In The Wild Using Improved CNN Supervisioncited by 494
  9. 2017Learning from Synthetic Humanscited by 359
  10. 20163D Menagerie: Modeling the 3D shape and pose of animalscited by 312
  11. 2017Integral Human Pose Regressioncited by 256
  12. 2018Learning to Estimate 3D Human Pose and Shape from a Single Color Imagecited by 226
28 more
  1. 2023Perpetual Humanoid Control for Real-time Simulated Avatarscited by 218
  2. 2021AGORA: Avatars in Geography Optimized for Regression Analysiscited by 205
  3. 2021MHFormer: Multi-Hypothesis Transformer for 3D Human Pose Estimationcited by 151
  4. 2017Unsupervised Learning of Disentangled Representations from Videocited by 139
  5. 2024Expressive Whole-Body Control for Humanoid Robotscited by 139
  6. 2021Human Pose Regression with Residual Log-likelihood Estimationcited by 131
  7. 2018Unsupervised Learning of Object Landmarks through Conditional Image Generationcited by 114
  8. 2020Synthetic Training for Accurate 3D Human Pose and Shape Estimation in the Wildcited by 82
  9. 2020Who Left the Dogs Out? 3D Animal Reconstruction with Expectation Maximization in the Loopcited by 81
  10. 2018Discovery of Latent 3D Keypoints via End-to-end Geometric Reasoningcited by 79
  11. 2016Stochastic Multiple Choice Learning for Training Diverse Deep Ensemblescited by 77
  12. 2018Unsupervised Discovery of Object Landmarks as Structural Representationscited by 73
  13. 2020CanonPose: Self-Supervised Monocular 3D Human Pose Estimation in the Wildcited by 65
  14. 2023SynBody: Synthetic Dataset with Layered Human Models for 3D Human Perception and Modelingcited by 54
  15. 2022BARC: Learning to Regress 3D Dog Shape from Images by Exploiting Breed Informationcited by 51
  16. 2020Self-Supervised 3D Human Pose Estimation via Part Guided Novel Image Synthesiscited by 44
  17. 2020Kinematic-Structure-Preserved Representation for Unsupervised 3D Human Pose Estimationcited by 26
  18. 2021Unsupervised Human Pose Estimation through Transforming Shape Templatescited by 26
  19. 2019Neural Scene Decomposition for Multi-Person Motion Capturecited by 21
  20. 2024SelfPose3d: Self-Supervised Multi-Person Multi-View 3d Pose Estimationcited by 21
  21. 2022AutoLink: Self-supervised Learning of Human Skeletons and Object Outlines by Linking Keypointscited by 17
  22. 2024Within the Dynamic Context: Inertia-aware 3D Human Modeling with Pose Sequencecited by 11
  23. 2022BKinD-3D: Self-Supervised 3D Keypoint Discovery from Multi-View Videoscited by 10
  24. 2023Self-supervised 3D Human Pose Estimation from a Single Imagecited by 8
  25. 2024Diffusion Models are Efficient Data Generators for Human Mesh Recoverycited by 8
  26. 2020Temporal Representation Learning on Monocular Videos for 3D Human Pose Estimationcited by 7
  27. 2023Improving 2D Human Pose Estimation in Rare Camera Views with Synthetic Datacited by 2
  28. 2023Mask as Supervision: Leveraging Unified Mask Information for Unsupervised 3D Pose Estimationcited by 2

Built on by

Not cited yet.