mlsanityData to 7 Oct 2026

Unsupervised learning of object landmarks by factorized spatial embeddings

arXiv 1705.02193 ↗First posted 5 May 2017cs.CV stat.MLcited by 64

Not on the map. Why?

Abstract (arXiv)

Learning automatically the structure of object categories remains an important open problem in computer vision. In this paper, we propose a novel unsupervised approach that can discover and learn landmarks in object categories, thus characterizing their structure. Our approach is based on factorizing image deformations, as induced by a viewpoint change or an object deformation, by learning a deep neural network that detects landmarks consistently with such visual effects. Furthermore, we show that the learned landmarks establish meaningful correspondences between different object instances in a category without having to impose this requirement explicitly. We assess the method qualitatively on a variety of object types, natural and man-made. We also show that our unsupervised landmarks are highly predictive of manually-annotated landmarks in face benchmark datasets, and can be used to regress these with a high degree of accuracy.

Key facts. Cited by 64 papers in the map (data through 7 Oct 2026). First posted on arXiv 5 May 2017.

Builds on

No reference list yet.

Built on by

Cited by year

Citing papers per year: 2021: 25; 2022: 14; 2023: 8; 2024: 9; 2025: 8.

64 papers.

Sort:

By area:

  1. 2021Deep ViT Features as Dense Visual Descriptorscited by 327
  2. 2023TAPIR: Tracking Any Point with per-frame Initialization and temporal Refinementcited by 281
  3. 2022TAP-Vid: A Benchmark for Tracking Any Point in a Videocited by 258
  4. 2021PiCIE: Unsupervised Semantic Segmentation using Invariance and Equivariance in Clusteringcited by 163
  5. 2021PartImageNet: A Large, High-Quality Dataset of Partscited by 107
  6. 2021Repurposing GANs for One-shot Semantic Part Segmentationcited by 88
  7. 2021Learning by Watching: Physical Imitation of Manipulation Skills from Human Videoscited by 87
  8. 2021Unsupervised Part Discovery from Contrastive Reconstructioncited by 57
  9. 2021KeypointDeformer: Unsupervised 3D Keypoint Discovery for Shape Controlcited by 50
  10. 2022Exploiting Transformation Invariance and Equivariance for Self-supervised Sound Localisationcited by 43
  11. 2021Unsupervised Part Segmentation through Disentangling Appearance and Shapecited by 29
  12. 2021Self-Supervised Keypoint Discovery in Behavioral Videoscited by 23
38 more
  1. 2021GANSeg: Learning to Segment by Unsupervised Hierarchical Image Generationcited by 19
  2. 2022Unsupervised Image Representation Learning with Deep Latent Particlescited by 19
  3. 2021Weakly supervised segmentation with cross-modality equivariant constraintscited by 18
  4. 2021A Novel Dataset for Keypoint Detection of quadruped Animals from Imagescited by 18
  5. 2022Part-based Face Recognition with Vision Transformerscited by 18
  6. 2023NCP: Neural Correspondence Prior for Effective Unsupervised Shape Matchingcited by 18
  7. 2022AutoLink: Self-supervised Learning of Human Skeletons and Object Outlines by Linking Keypointscited by 17
  8. 2023Unified Mask Embedding and Correspondence Learning for Self-Supervised Video Segmentationcited by 16
  9. 2021Semi-supervised Keypoint Localizationcited by 15
  10. 2023Unsupervised Keypoints from Pretrained Diffusion Modelscited by 15
  11. 2021Discovering Relationships between Object Categories via Universal Canonical Mapscited by 13
  12. 2024DualPM: Dual Posed-Canonical Point Maps for 3D Shape and Pose Reconstructioncited by 10
  13. 2022Demystifying Unsupervised Semantic Correspondence Estimationcited by 9
  14. 2023Few-shot Geometry-Aware Keypoint Localizationcited by 6
  15. 2024SCE-MAE: Selective Correspondence Enhancement with Masked Autoencoder for Self-Supervised Landmark Estimationcited by 6
  16. 2021LatentKeypointGAN: Controlling Images via Latent Keypointscited by 5
  17. 2021Semi-supervised Viewpoint Estimation with Geometry-aware Conditional Generationcited by 5
  18. 2021GaussiGAN: Controllable Image Synthesis with 3D Gaussians from Unposed Silhouettescited by 5
  19. 2021Weakly Supervised Keypoint Discoverycited by 5
  20. 2022TUSK: Task-Agnostic Unsupervised Keypointscited by 5
  21. 2021Landmarks Augmentation with Manifold-Barycentric Oversamplingcited by 4
  22. 2022From Keypoints to Object Landmarks via Self-Training Correspondence: A novel approach to Unsupervised Landmark Discoverycited by 4
  23. 2022Driving Points Prediction For Abdominal Probabilistic Registrationcited by 4
  24. 2024WPS-SAM: Towards Weakly-Supervised Part Segmentation with Foundation Modelscited by 4
  25. 2021JOKR: Joint Keypoint Representation for Unsupervised Cross-Domain Motion Retargetingcited by 3
  26. 2022Pose-MUM : Reinforcing Key Points Relationship for Semi-Supervised Human Pose Estimationcited by 3
  27. 2022Polymorphic-GAN: Generating Aligned Samples across Multiple Domains with Learned Morph Mapscited by 3
  28. 2023Bias-Compensated Integral Regression for Human Pose Estimationcited by 3
  29. 2024Pose-Guided Self-Training with Two-Stage Clustering for Unsupervised Landmark Discoverycited by 3
  30. 2024Pose Prior Learner: Unsupervised Categorical Prior Learning for Pose Estimationcited by 3
  31. 2021Pretrained equivariant features improve unsupervised landmark discoverycited by 2
  32. 2021Self-Supervised Object Detection via Generative Image Synthesiscited by 2
  33. 2022LEAD: Self-Supervised Landmark Estimation by Aligning Distributions of Feature Similaritycited by 2
  34. 2022Unsupervised 3D Keypoint Discovery with Multi-View Geometrycited by 2
  35. 2024Unsupervised Part Discovery via Dual Representation Alignmentcited by 2
  36. 2024Learning Keypoints for Multi-Agent Behavior Analysis using Self-Supervisioncited by 2
  37. 2025Towards Efficient Real-Time Video Motion Transfer via Generative Time Series Modelingcited by 2
  38. 2021Structure-Regularized Attention for Deformable Object Representationcited by 1

50 most cited of 64 All 64 (JSON)