mlsanityData to 7 Oct 2026

Rethinking the Heatmap Regression for Bottom-up Human Pose Estimation

arXiv 2012.15175 ↗First posted 30 Dec 2020cs.CVcited by 69

Not on the map. Why?

Abstract (arXiv)

Heatmap regression has become the most prevalent choice for nowadays human pose estimation methods. The ground-truth heatmaps are usually constructed via covering all skeletal keypoints by 2D gaussian kernels. The standard deviations of these kernels are fixed. However, for bottom-up methods, which need to handle a large variance of human scales and labeling ambiguities, the current practice seems unreasonable. To better cope with these problems, we propose the scale-adaptive heatmap regression (SAHR) method, which can adaptively adjust the standard deviation for each keypoint. In this way, SAHR is more tolerant of various human scales and labeling ambiguities. However, SAHR may aggravate the imbalance between fore-background samples, which potentially hurts the improvement of SAHR. Thus, we further introduce the weight-adaptive heatmap regression (WAHR) to help balance the fore-background samples. Extensive experiments show that SAHR together with WAHR largely improves the accuracy of bottom-up human pose estimation. As a result, we finally outperform the state-of-the-art model by +1.5AP and achieve 72.0AP on COCO test-dev2017, which is com-arable with the performances of most top-down methods. Source codes are available at https://github.com/greatlog/SWAHR-HumanPose.

Key facts. Cited by 69 papers in the map (data through 7 Oct 2026). First posted on arXiv 30 Dec 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 9; 2022: 13; 2023: 22; 2024: 13; 2025: 8; 2026: 4.

69 papers.

Sort:

By area:

  1. 2023RTMPose: Real-Time Multi-Person Pose Estimation based on MMPosecited by 170
  2. 2024DINO-X: A Unified Vision Model for Open-World Object Detection and Understandingcited by 85
  3. 2021SimCC: a Simple Coordinate Classification Perspective for Human Pose Estimationcited by 68
  4. 2023Explicit Box Detection Unifies End-to-End Multi-Person Pose Estimationcited by 48
  5. 2023Human Pose as Compositional Tokenscited by 45
  6. 2022Poseur: Direct Human Pose Regression with Transformerscited by 43
  7. 2021TFPose: Direct Human Pose Estimation with Transformerscited by 42
  8. 2021Polarized Self-Attention: Towards High-quality Pixel-wise Regressioncited by 39
  9. 2023RTMO: Towards High-Performance One-Stage Real-Time Multi-Person Pose Estimationcited by 33
  10. 2021Few-shot Keypoint Detection with Uncertainty Learning for Unseen Speciescited by 27
  11. 2023Distribution-Aligned Diffusion for Human Mesh Recoverycited by 27
  12. 2023Group Pose: A Simple Baseline for End-to-End Multi-person Pose Estimationcited by 26
38 more
  1. 2023MolGrapher: Graph-based Visual Recognition of Chemical Structurescited by 24
  2. 20222D Human Pose Estimation: A Surveycited by 21
  3. 2021Rethinking Keypoint Representations: Modeling Keypoints and Poses as Objects for Multi-Person Human Pose Estimationcited by 20
  4. 2022Learnable human mesh triangulation for 3D human pose and shape estimationcited by 16
  5. 2024GenMix: Effective Data Augmentation with Generative Diffusion Model Image Editingcited by 14
  6. 2022Dual networks based 3D Multi-Person Pose Estimation from Monocular Videocited by 13
  7. 2022QueryPose: Sparse Multi-Person Pose Regression via Spatial-Aware Part-Level Querycited by 13
  8. 2023Neural Interactive Keypoint Detectioncited by 12
  9. 2024Video-Based Human Pose Regression via Decoupled Space-Time Aggregationcited by 10
  10. 2022Heatmap Distribution Matching for Human Pose Estimationcited by 9
  11. 2023You Only Learn One Query: Learning Unified Human Query for Single-Stage Multi-Person Multi-Task Human-Centric Perceptioncited by 9
  12. 2021Relation-Based Associative Joint Location for Human Pose Estimation in Videoscited by 7
  13. 2023Self-Correctable and Adaptable Inference for Generalizable Human Pose Estimationcited by 7
  14. 2024Meta-Point Learning and Refining for Category-Agnostic Pose Estimationcited by 7
  15. 2021AdaptivePose: Human Parts as Adaptive Pointscited by 6
  16. 2022Learning Quality-aware Representation for Multi-person Pose Regressioncited by 6
  17. 2022AdaptivePose++: A Powerful Single-Stage Network for Multi-Person Pose Regressioncited by 6
  18. 2023Topology Repairing of Disconnected Pulmonary Airways and Vessels: Baselines and a Datasetcited by 6
  19. 2021Probabilistic Spatial Analysis in Quantitative Microscopy with Uncertainty-Aware Cell Detection using Deep Bayesian Regression of Density Mapscited by 5
  20. 2022Bottom-Up 2D Pose Estimation via Dual Anatomical Centers for Small-Scale Personscited by 5
  21. 2025MoFM: A Large-Scale Human Motion Foundation Modelcited by 5
  22. 2022Self-Constrained Inference Optimization on Structural Groups for Human Pose Estimationcited by 4
  23. 2023LAMP: Leveraging Language Prompts for Multi-person Pose Estimationcited by 4
  24. 2024Human Modelling and Pose Estimation Overviewcited by 4
  25. 20222D Human Pose Estimation with Explicit Anatomical Keypoints Structure Constraintscited by 3
  26. 2023Bias-Compensated Integral Regression for Human Pose Estimationcited by 3
  27. 2023Deep Single Image Camera Calibration by Heatmap Regression to Recover Fisheye Images Under Manhattan World Assumptioncited by 3
  28. 2024Rethinking Self-training for Semi-supervised Landmark Detection: A Selection-free Approachcited by 3
  29. 2026From Offline to Periodic Adaptation for Pose-Based Shoplifting Detection in Real-world Retail Securitycited by 3
  30. 2022Adversarial Focal Loss: Asking Your Discriminator for Hard Examplescited by 2
  31. 2023Joint Coordinate Regression and Association For Multi-Person Pose Estimation, A Pure Neural Network Approachcited by 2
  32. 2023BoIR: Box-Supervised Instance Representation for Multi-Person Pose Estimationcited by 2
  33. 2023MD-IQA: Learning Multi-scale Distributed Image Quality Assessment with Semi Supervised Learning for Low Dose CTcited by 2
  34. 2024DHRNet: A Dual-Path Hierarchical Relation Network for Multi-Person Pose Estimationcited by 2
  35. 2024Recoverable Anonymization for Pose Estimation: A Privacy-Enhancing Approachcited by 2
  36. 2024V-Hands: Touchscreen-based Hand Tracking for Remote Whiteboard Interactioncited by 2
  37. 2024CapeLLM: Support-Free Category-Agnostic Pose Estimation with Multimodal Large Language Modelscited by 2
  38. 2024ReMP: Reusable Motion Prior for Multi-domain 3D Human Pose Estimation and Motion Inbetweeningcited by 2

50 most cited of 69 All 69 (JSON)