mlsanityData to 7 Oct 2026

PoseFormerV2: Exploring Frequency Domain for Efficient and Robust 3D Human Pose Estimation

arXiv 2303.17472 ↗First posted 30 Mar 2023cs.CVcited by 91

Not on the map. Why?

Abstract (arXiv)

Recently, transformer-based methods have gained significant success in sequential 2D-to-3D lifting human pose estimation. As a pioneering work, PoseFormer captures spatial relations of human joints in each video frame and human dynamics across frames with cascaded transformer layers and has achieved impressive performance. However, in real scenarios, the performance of PoseFormer and its follow-ups is limited by two factors: (a) The length of the input joint sequence; (b) The quality of 2D joint detection. Existing methods typically apply self-attention to all frames of the input sequence, causing a huge computational burden when the frame number is increased to obtain advanced estimation accuracy, and they are not robust to noise naturally brought by the limited capability of 2D joint detectors. In this paper, we propose PoseFormerV2, which exploits a compact representation of lengthy skeleton sequences in the frequency domain to efficiently scale up the receptive field and boost robustness to noisy 2D joint detection. With minimum modifications to PoseFormer, the proposed method effectively fuses features both in the time domain and frequency domain, enjoying a better speed-accuracy trade-off than its precursor. Extensive experiments on two benchmark datasets (i.e., Human3.6M and MPI-INF-3DHP) demonstrate that the proposed approach significantly outperforms the original PoseFormer and other transformer-based variants. Code is released at urlhttps://github.com/QitaoZhao/PoseFormerV2.

Key facts. Cited by 91 papers in the map (data through 7 Oct 2026). Builds on 38 of its 49 references that are in the map. First posted on arXiv 30 Mar 2023.

Builds on

38 of 49 references.

References · 38

  1. 2017Attention Is All You Needcited by 59,932
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  3. 2020Language Models are Few-Shot Learnerscited by 32,199
  4. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  5. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,082
  6. 2020End-to-End Object Detection with Transformerscited by 7,384
  7. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,339
  8. 2020Training data-efficient image transformers & distillation through attentioncited by 4,560
  9. 2020Deformable DETR: Deformable Transformers for End-to-End Object Detectioncited by 3,410
  10. 2021ViViT: A Video Vision Transformercited by 1,610
  11. 2021Is Space-Time Attention All You Need for Video Understanding?cited by 1,609
  12. 2019Deep High-Resolution Representation Learning for Human Pose Estimationcited by 1,339
26 more
  1. 2021SimMIM: A Simple Framework for Masked Image Modelingcited by 1,147
  2. 2016Stacked Hourglass Networks for Human Pose Estimationcited by 913
  3. 2021Masked Feature Prediction for Self-Supervised Visual Pre-Trainingcited by 573
  4. 2016Monocular 3D Human Pose Estimation In The Wild Using Improved CNN Supervisioncited by 493
  5. 20183D human pose estimation in video with temporal convolutions and semi-supervised trainingcited by 401
  6. 2017Cascaded Pyramid Network for Multi-Person Pose Estimationcited by 341
  7. 2020I2L-MeshNet: Image-to-Lixel Prediction Network for Accurate 3D Human Pose and Mesh Estimation from a Single RGB Imagecited by 274
  8. 20213D Human Pose Estimation with Spatial and Temporal Transformerscited by 271
  9. 2021Global Filter Networks for Image Classificationcited by 257
  10. 2019Learning Trajectory Dependencies for Human Motion Predictioncited by 245
  11. 2021Adaptive Fourier Neural Operators: Efficient Token Mixers for Transformerscited by 237
  12. 2020Deep Learning-Based Human Pose Estimation: A Surveycited by 205
  13. 2020Learning in the Frequency Domaincited by 196
  14. 2020History Repeats Itself: Human Motion Prediction via Motion Attentioncited by 164
  15. 2022MixSTE: Seq2seq Mixed Spatio-Temporal Encoder for 3D Human Pose Estimation in Videocited by 158
  16. 2021MHFormer: Multi-Hypothesis Transformer for 3D Human Pose Estimationcited by 151
  17. 2022Multiview Transformers for Video Recognitioncited by 148
  18. 2017VNect: Real-time 3D Human Pose Estimation with a Single RGB Cameracited by 129
  19. 2018Ordinal Depth Supervision for 3D Human Pose Estimationcited by 107
  20. 2020SRNet: Improving Generalization in 3D Human Pose Estimation with a Split-and-Recombine Approachcited by 94
  21. 2021Exploiting Temporal Contexts with Strided Transformer for 3D Human Pose Estimationcited by 94
  22. 2020Motion Guided 3D Pose Estimation from Videoscited by 85
  23. 2020Cascaded deep monocular 3D human pose estimation with evolutionary training datacited by 80
  24. 2022P-STMO: Pre-Trained Spatial Temporal Many-to-One Model for 3D Human Pose Estimationcited by 80
  25. 2022Uplift and Upsample: Efficient 3D Human Pose Estimation with Uplifting Transformerscited by 33
  26. 2019Trajectory Space Factorization for Deep Video-Based 3D Human Pose Estimationcited by 30

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 14; 2024: 25; 2025: 35; 2026: 17.

91 papers.

Sort:

By area:

  1. 2023MotionAGFormer: Enhancing 3D Human Pose Estimation with a Transformer-GCNFormer Networkcited by 62
  2. 2023EmotionGesture: Audio-Driven Diverse Emotional Co-Speech 3D Gesture Generationcited by 40
  3. 2024KTPFormer: Kinematics and Trajectory Prior Knowledge-Enhanced Transformer for 3D Human Pose Estimationcited by 33
  4. 2024Multi-agent Long-term 3D Human Pose Forecasting via Interaction-aware Trajectory Conditioningcited by 29
  5. 2024PhysPT: Physics-aware Pretrained Transformer for Estimating Human Dynamics from Monocular Videoscited by 27
  6. 20233D Human Pose Perception from Egocentric Stereo Videoscited by 23
  7. 2024FinePOSE: Fine-Grained Prompt-Driven 3D Human Pose Estimation via Diffusion Modelscited by 23
  8. 2023Hourglass Tokenizer for Efficient Transformer-Based 3D Human Pose Estimationcited by 15
  9. 2025AthletePose3D: A Benchmark Dataset for 3D Human Pose Estimation and Kinematic Validation in Athletic Movementscited by 14
  10. 2024Attention-Propagation Network for Egocentric Heatmap to 3D Pose Liftingcited by 13
  11. 2025CoMotion: Concurrent Multi-person 3D Motioncited by 13
  12. 2025Learning Dense Hand Contact Estimation from Imbalanced Datacited by 12
38 more
  1. 2023A Single 2D Pose with Context is Worth Hundreds for 3D Human Pose Estimationcited by 11
  2. 2023FreeMan: Towards Benchmarking 3D Human Pose Estimation under Real-World Conditionscited by 7
  3. 2025Bring Your Rear Cameras for Egocentric 3D Human Pose Estimationcited by 7
  4. 2025GAITGen: Disentangled Motion-Pathology Impaired Gait Generative Model -- Bringing Motion Generation to the Clinical Domaincited by 7
  5. 2023Human Pose-based Estimation, Tracking and Action Recognition with Deep Learning: A Surveycited by 6
  6. 2024HumMUSS: Human Motion Understanding using State Space Modelscited by 6
  7. 2024Benchmarking Skeleton-based Motion Encoder Models for Clinical Applications: Estimating Parkinson's Disease Severity in Walking Sequencescited by 6
  8. 2025Exploring Temporal Dynamics in Event-based Eye Trackercited by 6
  9. 2025Frequency-Semantic Enhanced Variational Autoencoder for Zero-Shot Skeleton-based Action Recognitioncited by 6
  10. 2025H2OT: Hierarchical Hourglass Tokenizer for Efficient Video Pose Transformerscited by 5
  11. 2025CARE-PD: A Multi-Site Anonymized Clinical Dataset for Parkinson's Disease Gait Assessmentcited by 5
  12. 2023MPM: A Unified 2D-3D Human Pose Representation via Masked Pose Modelingcited by 4
  13. 2023LAMP: Leveraging Language Prompts for Multi-person Pose Estimationcited by 4
  14. 2023Double-chain Constraints for 3D Human Pose Estimation in Images and Videoscited by 4
  15. 2024EgoPoseFormer: A Simple Baseline for Stereo Egocentric 3D Human Pose Estimationcited by 4
  16. 2025HGMamba: Enhancing 3D Human Pose Estimation with a HyperGCN-Mamba Networkcited by 4
  17. 2025FastHMR: Accelerating Human Mesh Recovery via Token and Layer Merging with Diffusion Decodingcited by 4
  18. 2026THFM: A Unified Video Foundation Model for 4D Human Perception and Beyondcited by 4
  19. 2024PADS: Plug-and-Play 3D Human Pose Analysis via Diffusion Generative Modelingcited by 3
  20. 2024Enhancing 3D Human Pose Estimation Amidst Severe Occlusion with Dual Transformer Fusioncited by 3
  21. 2024Optimizing Local-Global Dependencies for Accurate 3D Human Pose Estimationcited by 3
  22. 2025AGIR: Assessing 3D Gait Impairment with Reasoning based on LLMscited by 3
  23. 2025A Structure-aware and Motion-adaptive Framework for 3D Human Pose Estimation with Mambacited by 3
  24. 2026QuaMo: Quaternion Motions for Vision-based 3D Human Kinematics Capturecited by 3
  25. 2023STRIDE: Single-video based Temporally Continuous Occlusion-Robust 3D Pose Estimationcited by 2
  26. 20243D Human Pose Estimation with Occlusions: Introducing BlendMimic3D Dataset and GCN Refinementcited by 2
  27. 2024Di2Pose: Discrete Diffusion Model for Occluded 3D Human Pose Estimationcited by 2
  28. 2024STGFormer: Spatio-Temporal GraphFormer for 3D Human Pose Estimation in Videocited by 2
  29. 2024SPiKE: 3D Human Pose from Point Cloud Sequencescited by 2
  30. 2024BLAPose: Enhancing 3D Human Pose Estimation with Bone Length Adjustmentcited by 2
  31. 2024ConvMixFormer- A Resource-efficient Convolution Mixer for Transformer-based Dynamic Hand Gesture Recognitioncited by 2
  32. 2025NanoHTNet: Nano Human Topology Network for Efficient 3D Human Pose Estimationcited by 2
  33. 2025SCJD: Sparse Correlation and Joint Distillation for Efficient 3D Human Pose Estimationcited by 2
  34. 2025Benchmarking 3D Human Pose Estimation Models under Occlusionscited by 2
  35. 2025PoseBench3D: A Cross-Dataset Analysis Framework for 3D Human Pose Estimation via Pose Lifting Networkscited by 2
  36. 2025Mamba-Driven Topology Fusion for Monocular 3D Human Pose Estimationcited by 2
  37. 2025PUMPS: Skeleton-Agnostic Point-based Universal Motion Pre-Training for Synthesis in Human Motion Taskscited by 2
  38. 2025DreamPose3D: Hallucinative Diffusion with Prompt Learning for 3D Human Pose Estimationcited by 2

50 most cited of 91 All 91 (JSON)