mlsanityData to 14 Oct 2026

DeciWatch: A Simple Baseline for 10x Efficient 2D and 3D Pose Estimation

arXiv 2203.08713 ↗First posted 16 Mar 2022cs.CVcited by 29

Not on the map. Why?

Abstract (arXiv)

This paper proposes a simple baseline framework for video-based 2D/3D human pose estimation that can achieve 10 times efficiency improvement over existing works without any performance degradation, named DeciWatch. Unlike current solutions that estimate each frame in a video, DeciWatch introduces a simple yet effective sample-denoise-recover framework that only watches sparsely sampled frames, taking advantage of the continuity of human motions and the lightweight pose representation. Specifically, DeciWatch uniformly samples less than 10% video frames for detailed estimation, denoises the estimated 2D/3D poses with an efficient Transformer architecture, and then accurately recovers the rest of the frames using another Transformer-based network. Comprehensive experimental results on three video-based human pose estimation and body mesh recovery tasks with four datasets validate the efficiency and effectiveness of DeciWatch. Code is available at https://github.com/cure-lab/DeciWatch.

Key facts. Cited by 29 papers in the map (data through 14 Oct 2026). Builds on 43 of its 65 references that are in the map. First posted on arXiv 16 Mar 2022.

Builds on

43 of 65 references.

References · 43

  1. 2017Attention Is All You Needcited by 60,361
  2. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  3. 2015Distilling the Knowledge in a Neural Networkcited by 10,324
  4. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 6,013
  5. 2019Deep High-Resolution Representation Learning for Human Pose Estimationcited by 1,339
  6. 2018OpenPose: Realtime Multi-Person 2D Pose Estimation using Part Affinity Fieldscited by 1,162
  7. 2016Stacked Hourglass Networks for Human Pose Estimationcited by 915
  8. 2017End-to-end Recovery of Human Shape and Posecited by 885
  9. 2019Learning to Reconstruct 3D Human Pose and Shape via Model-fitting in the Loopcited by 627
  10. 2018Simple Baselines for Human Pose Estimation and Trackingcited by 502
  11. 2021AI Choreographer: Music Conditioned 3D Dance Generation with AIST++cited by 485
  12. 20183D human pose estimation in video with temporal convolutions and semi-supervised trainingcited by 403
31 more
  1. 2017A simple yet effective baseline for 3d human pose estimationcited by 401
  2. 2021PARE: Part Attention Regressor for 3D Human Body Estimationcited by 337
  3. 20213D Human Pose Estimation with Spatial and Temporal Transformerscited by 272
  4. 2021Robust Motion In-betweeningcited by 229
  5. 2020Deep Learning-Based Human Pose Estimation: A Surveycited by 207
  6. 2021GLAMR: Global Occlusion-Aware Human Mesh Recovery with Dynamic Camerascited by 153
  7. 2021Human Pose Regression with Residual Log-likelihood Estimationcited by 131
  8. 2018Human Motion Prediction via Spatio-Temporal Inpaintingcited by 114
  9. 2020Exemplar Fine-Tuning for 3D Human Model Fitting Towards In-the-Wild 3D Human Pose Estimationcited by 113
  10. 2020SRNet: Improving Generalization in 3D Human Pose Estimation with a Split-and-Recombine Approachcited by 94
  11. 2021Lite-HRNet: A Lightweight High-Resolution Networkcited by 91
  12. 2020Convolutional Autoencoders for Human Motion Infillingcited by 73
  13. 2021SmoothNet: A Plug-and-Play Network for Refining Human Poses in Videoscited by 64
  14. 2021Learning Skeletal Graph Neural Networks for Hard 3D Pose Estimationcited by 62
  15. 2021Single-Shot Motion Completion with Transformercited by 59
  16. 2021Online Knowledge Distillation for Efficient Pose Estimationcited by 55
  17. 2018Recurrent Transition Networks for Character Locomotioncited by 52
  18. 2021Recent Advances in Monocular 2D and 3D Human Pose Estimation: A Deep Learning Perspectivecited by 50
  19. 2018Real-time 2D Multi-Person Pose Estimation on CPU: Lightweight OpenPosecited by 42
  20. 2019Dynamic Kernel Distillation for Efficient Pose Estimation in Videoscited by 30
  21. 2017LSTM Pose Machinescited by 28
  22. 2020A review of 3D human pose estimation algorithms for markerless motion capturecited by 24
  23. 2021A Lightweight Graph Transformer Network for Human Mesh Reconstruction from 2D Human Posecited by 21
  24. 2020Key Frame Proposal Network for Efficient Pose Estimation in Videoscited by 12
  25. 2021Part-Aware Measurement for Robust Multi-View Multi-Human 3D Pose Estimation and Trackingcited by 12
  26. 2019Simple and Lightweight Human Pose Estimationcited by 10
  27. 2020Lightweight 3D Human Pose Estimation Network Training Using Teacher-Student Learningcited by 7
  28. 2021Exploring Versatile Prior for Human Motion via Motion Frequency Guidancecited by 5
  29. 2018A Neural Network Approach to Missing Marker Reconstruction in Human Motion Capturecited by 2
  30. 2021Render In-between: Motion Guided Video Synthesis for Action Interpolationcited by 2
  31. 2021FasterPose: A Faster Simple Baseline for Human Pose Estimationcited by 1

Built on by

Cited by year

Citing papers per year: 2022: 4; 2023: 14; 2024: 4; 2025: 7.

29 papers.

Sort:

By area:

  1. 2022Are Transformers Effective for Time Series Forecasting?cited by 1,424
  2. 2023Motion-X: A Large-scale 3D Expressive Whole-body Human Motion Datasetcited by 227
  3. 2023One-Stage 3D Whole-Body Mesh Recovery with Component Aware Transformercited by 130
  4. 2023HumanSD: A Native Skeleton-Guided Diffusion Model for Human Image Generationcited by 105
  5. 2023DreamWaltz: Make a Scene with Complex 3D Animatable Avatarscited by 82
  6. 2023Human-Art: A Versatile Human-Centric Dataset Bridging Natural and Artificial Scenescited by 68
  7. 2023SynBody: Synthetic Dataset with Layered Human Models for 3D Human Perception and Modelingcited by 54
  8. 2024AiOS: All-in-One-Stage Expressive Human Pose and Shape Estimationcited by 42
  9. 2023Zolly: Zoom Focal Length Correctly for Perspective-Distorted Human Mesh Reconstructioncited by 41
  10. 2023Cyclic Test-Time Adaptation on Monocular Video for 3D Human Mesh Reconstructioncited by 18
  11. 2023Hourglass Tokenizer for Efficient Transformer-Based 3D Human Pose Estimationcited by 15
  12. 2023Neural Interactive Keypoint Detectioncited by 12
17 more
  1. 2024DreamWaltz-G: Expressive 3D Gaussian Avatars from Skeleton-Guided 2D Diffusioncited by 12
  2. 2023A Single 2D Pose with Context is Worth Hundreds for 3D Human Pose Estimationcited by 11
  3. 2024Deep learning for 3D human pose estimation and mesh recovery: A surveycited by 11
  4. 2023Human Pose-based Estimation, Tracking and Action Recognition with Deep Learning: A Surveycited by 6
  5. 2023Coordinate Transformer: Achieving Single-stage Multi-person Mesh Recovery from Videoscited by 5
  6. 2025H2OT: Hierarchical Hourglass Tokenizer for Efficient Video Pose Transformerscited by 5
  7. 2022Kinematic-aware Hierarchical Attention Network for Human Pose Estimation in Videoscited by 4
  8. 2023Active Transfer Learning for Efficient Video-Specific Human Pose Estimationcited by 4
  9. 2025AnyMoLe: Any Character Motion In-betweening Leveraging Video Diffusion Modelscited by 4
  10. 2025High-Resolution Spatiotemporal Modeling with Global-Local State Space Models for Video-Based Human Pose Estimationcited by 4
  11. 2022BlanketSet -- A clinical real-world in-bed action recognition and qualitative semi-synchronised MoCap datasetcited by 2
  12. 2025NanoHTNet: Nano Human Topology Network for Efficient 3D Human Pose Estimationcited by 2
  13. 2025Mamba-Driven Topology Fusion for Monocular 3D Human Pose Estimationcited by 2
  14. 2022BlanketGen - A synthetic blanket occlusion augmentation pipeline for MoCap datasetscited by 1
  15. 2025Spectral Compression Transformer with Line Pose Graph for Monocular 3D Human Pose Estimationcited by 1
  16. 2024RemoCap: Disentangled Representation Learning for Motion Capturecited by 0
  17. 2025Efficient Diffusion-Based 3D Human Pose Estimation with Hierarchical Temporal Pruningcited by 0