mlsanityData to 7 Oct 2026

Efficient, Self-Supervised Human Pose Estimation with Inductive Prior Tuning

arXiv 2311.02815 ↗First posted 6 Nov 2023cs.CVcited by 1

Not on the map. Why?

Abstract (arXiv)

The goal of 2D human pose estimation (HPE) is to localize anatomical landmarks, given an image of a person in a pose. SOTA techniques make use of thousands of labeled figures (finetuning transformers or training deep CNNs), acquired using labor-intensive crowdsourcing. On the other hand, self-supervised methods re-frame the HPE task as a reconstruction problem, enabling them to leverage the vast amount of unlabeled visual data, though at the present cost of accuracy. In this work, we explore ways to improve self-supervised HPE. We (1) analyze the relationship between reconstruction quality and pose estimation accuracy, (2) develop a model pipeline that outperforms the baseline which inspired our work, using less than one-third the amount of training data, and (3) offer a new metric suitable for self-supervised settings that measures the consistency of predicted body part length proportions. We show that a combination of well-engineered reconstruction losses and inductive priors can help coordinate pose learning alongside reconstruction in a self-supervised paradigm.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 33 of its 47 references that are in the map. First posted on arXiv 6 Nov 2023.

Builds on

33 of 47 references.

References · 33

  1. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  2. 2016Perceptual Losses for Real-Time Style Transfer and Super-Resolutioncited by 2,678
  3. 2016Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Networkcited by 1,898
  4. 2022ViTPose: Simple Vision Transformer Baselines for Human Pose Estimationcited by 450
  5. 2020Deblurring by Realistic Blurringcited by 187
  6. 2020Feature-metric Loss for Self-supervised Learning of Depth and Egomotioncited by 140
  7. 2019Self-supervised 6D Object Pose Estimation for Robot Manipulationcited by 100
  8. 2019Self-Supervised Learning of 3D Human Pose using Multi-view Geometrycited by 98
  9. 2016Learning Detailed Face Reconstruction from a Single Imagecited by 71
  10. 2020CanonPose: Self-Supervised Monocular 3D Human Pose Estimation in the Wildcited by 65
  11. 2023Human Pose as Compositional Tokenscited by 45
  12. 2020Self-Supervised 3D Human Pose Estimation via Part Guided Novel Image Synthesiscited by 44
21 more
  1. 2018Real-time 2D Multi-Person Pose Estimation on CPU: Lightweight OpenPosecited by 42
  2. 2015Learning to Generate Images with Perceptual Similarity Metricscited by 39
  3. 2021Multi-modal 3D Human Pose Estimation with 2D Weak Supervision in Autonomous Drivingcited by 39
  4. 2021Context Modeling in 3D Human Pose Estimation: A Unified Perspectivecited by 37
  5. 2022Uplift and Upsample: Efficient 3D Human Pose Estimation with Uplifting Transformerscited by 33
  6. 2022HuPR: A Benchmark for Human Pose Estimation Using Millimeter Wave Radarcited by 31
  7. 2021Unsupervised Human Pose Estimation through Transforming Shape Templatescited by 26
  8. 2022Uncertainty-Aware Adaptation for Self-Supervised 3D Human Pose Estimationcited by 25
  9. 2022HUM3DIL: Semi-supervised Multi-modal 3D Human Pose Estimation for Autonomous Drivingcited by 22
  10. 2020Self-supervised Learning for Precise Pick-and-place without Object Modelcited by 20
  11. 2021A Study of the Generalizability of Self-Supervised Representationscited by 18
  12. 2022Neural Groundplans: Persistent Neural Scene Representations from a Single Imagecited by 17
  13. 2019Simple and Lightweight Human Pose Estimationcited by 10
  14. 2021Self-Supervised 3D Human Pose Estimation with Multiple-View Geometrycited by 10
  15. 2020Attention! A Lightweight 2D Hand Pose Estimation Approachcited by 9
  16. 2023Neuromorphic High-Frequency 3D Dancing Pose Estimation in Dynamic Environmentcited by 8
  17. 2023Self-supervised 3D Human Pose Estimation from a Single Imagecited by 8
  18. 2023Weakly Supervised Multi-Modal 3D Human Body Pose Estimation for Autonomous Drivingcited by 8
  19. 2021In-Bed Human Pose Estimation from Unseen and Privacy-Preserving Image Domainscited by 7
  20. 2023CameraPose: Weakly-Supervised Monocular 3D Human Pose Estimation by Leveraging In-the-wild 2D Annotationscited by 6
  21. 2022DPIT: Dual-Pipeline Integrated Transformer for Human Pose Estimationcited by 3

Built on by

Cited by year

Citing papers per year: 2024: 1.

1 paper.

  1. 2024Pose Prior Learner: Unsupervised Categorical Prior Learning for Pose Estimationcited by 3