mlsanityData to 7 Oct 2026

Recurrent Feature Mining and Keypoint Mixup Padding for Category-Agnostic Pose Estimation

arXiv 2503.21140 ↗First posted 27 Mar 2025cs.CVcited by 2

Not on the map. Why?

Abstract (arXiv)

Category-agnostic pose estimation aims to locate keypoints on query images according to a few annotated support images for arbitrary novel classes. Existing methods generally extract support features via heatmap pooling, and obtain interacted features from support and query via cross-attention. Hence, these works neglect to mine fine-grained and structure-aware (FGSA) features from both support and query images, which are crucial for pixel-level keypoint localization. To this end, we propose a novel yet concise framework, which recurrently mines FGSA features from both support and query images. Specifically, we design a FGSA mining module based on deformable attention mechanism. On the one hand, we mine fine-grained features by applying deformable attention head over multi-scale feature maps. On the other hand, we mine structure-aware features by offsetting the reference points of keypoints to their linked keypoints. By means of above module, we recurrently mine FGSA features from support and query images, and thus obtain better support features and query estimations. In addition, we propose to use mixup keypoints to pad various classes to a unified keypoint number, which could provide richer supervision than the zero padding used in existing works. We conduct extensive experiments and in-depth studies on large-scale MP-100 dataset, and outperform SOTA method dramatically (+3.2%[email protected]). Code is avaiable at https://github.com/chenbys/FMMP.

Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 45 of its 64 references that are in the map. First posted on arXiv 27 Mar 2025.

Builds on

45 of 64 references.

References · 45

  1. 2017Attention Is All You Needcited by 59,932
  2. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  3. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  4. 2020End-to-End Object Detection with Transformerscited by 7,384
  5. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,155
  6. 2017Squeeze-and-Excitation Networkscited by 4,763
  7. 2020Deformable DETR: Deformable Transformers for End-to-End Object Detectioncited by 3,410
  8. 2018CBAM: Convolutional Block Attention Modulecited by 2,463
  9. 2017Deformable Convolutional Networkscited by 1,481
  10. 2015Spatial Transformer Networkscited by 1,423
  11. 2018Dual Attention Network for Scene Segmentationcited by 1,001
  12. 2018Deformable ConvNets v2: More Deformable, Better Resultscited by 666
33 more
  1. 2018DensePose: Dense Human Pose Estimation In The Wildcited by 573
  2. 2017Residual Attention Network for Image Classificationcited by 385
  3. 2017Cascaded Pyramid Network for Multi-Person Pose Estimationcited by 341
  4. 2019HigherHRNet: Scale-Aware Representation Learning for Bottom-Up Human Pose Estimationcited by 247
  5. 2018CrowdPose: Efficient Crowded Scenes Pose Estimation and A New Benchmarkcited by 214
  6. 2019Distribution-Aware Coordinate Representation for Human Pose Estimationcited by 165
  7. 2021Human Pose Regression with Residual Log-likelihood Estimationcited by 131
  8. 2019Cross-Domain Adaptation for Animal Pose Estimationcited by 108
  9. 2019Single-Stage Multi-Person Pose Machinescited by 95
  10. 2018Learn To Pay Attentioncited by 83
  11. 2020Rethinking the Heatmap Regression for Bottom-up Human Pose Estimationcited by 69
  12. 2018ApolloCar3D: A Large 3D Car Instance Understanding Benchmark for Autonomous Drivingcited by 59
  13. 2022Poseur: Direct Human Pose Regression with Transformerscited by 43
  14. 2023Relax Image-Specific Prompt Requirement in SAM: A Single Generic Prompt for Segmenting Camouflaged Objectscited by 40
  15. 2022Pose for Everything: Towards Category-Agnostic Pose Estimationcited by 36
  16. 20193FabRec: Fast Few-shot Face alignment by Reconstructioncited by 33
  17. 2021Few-shot Keypoint Detection with Uncertainty Learning for Unseen Speciescited by 27
  18. 2022CLAMP: Prompt-based Contrastive Learning for Connecting Language and Animal Posecited by 25
  19. 2024Multi-Patch Prediction: Adapting LLMs for Time Series Representation Learningcited by 25
  20. 2023X-Pose: Detecting Any Keypointscited by 21
  21. 2020Weak-shot Fine-grained Classification via Similarity Transfercited by 18
  22. 2023A Graph-Based Approach for Category-Agnostic Pose Estimationcited by 16
  23. 2022Weak-shot Semantic Segmentation via Dual Similarity Transfercited by 10
  24. 2023Open-Vocabulary Animal Keypoint Detection with Semantic-feature Matchingcited by 10
  25. 2024CapeX: Category-Agnostic Pose Estimation from Textual Point Explanationcited by 9
  26. 2023From Saliency to DINO: Saliency-guided Vision Transformer for Few-shot Keypoint Detectioncited by 8
  27. 2024OpenKD: Opening Prompt Diversity for Zero- and Few-shot Keypoint Detectioncited by 8
  28. 2023UniAP: Towards Universal Animal Perception in Vision via Few-shot Learningcited by 7
  29. 2024Meta-Point Learning and Refining for Category-Agnostic Pose Estimationcited by 7
  30. 2021MetaCloth: Learning Unseen Tasks of Dense Fashion Landmark Detection from a Few Samplescited by 6
  31. 2021Few-Shot Model Adaptation for Customized Facial Landmark Detection, Segmentation, Stylization and Shadow Removalcited by 5
  32. 2024KptLLM: Unveiling the Power of Large Language Model for Keypoint Comprehensioncited by 5
  33. 2024SCAPE: A Simple and Strong Category-Agnostic Pose Estimatorcited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 1; 2026: 1.

2 papers.

  1. 2025Talking Points: Describing and Localizing Pixelscited by 0
  2. 2026GenCape: Structure-Inductive Generative Modeling for Category-Agnostic Pose Estimationcited by 0