mlsanityData to 7 Oct 2026

Unsupervised Discovery of the Long-Tail in Instance Segmentation Using Hierarchical Self-Supervision

arXiv 2104.01257 ↗First posted 2 Apr 2021cs.CVcited by 39

Not on the map. Why?

Abstract (arXiv)

Instance segmentation is an active topic in computer vision that is usually solved by using supervised learning approaches over very large datasets composed of object level masks. Obtaining such a dataset for any new domain can be very expensive and time-consuming. In addition, models trained on certain annotated categories do not generalize well to unseen objects. The goal of this paper is to propose a method that can perform unsupervised discovery of long-tail categories in instance segmentation, through learning instance embeddings of masked regions. Leveraging rich relationship and hierarchical structure between objects in the images, we propose self-supervised losses for learning mask embeddings. Trained on COCO dataset without additional annotations of the long-tail objects, our model is able to discover novel and more fine-grained objects than the common categories in COCO. We show that the model achieves competitive quantitative results on LVIS as compared to the supervised and partially supervised methods.

Key facts. Cited by 39 papers in the map (data through 7 Oct 2026). Builds on 36 of its 49 references that are in the map. First posted on arXiv 2 Apr 2021.

Builds on

36 of 49 references.

References · 36

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  4. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  5. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,441
  6. 2017Mask R-CNNcited by 7,259
  7. 2013Auto-Encoding Variational Bayescited by 3,980
  8. 2016Context Encoders: Feature Learning by Inpaintingcited by 1,424
  9. 2019LVIS: A Dataset for Large Vocabulary Instance Segmentationcited by 1,215
  10. 2016Colorful Image Colorizationcited by 1,113
  11. 2015Unsupervised Visual Representation Learning by Context Predictioncited by 1,062
  12. 2018Path Aggregation Network for Instance Segmentationcited by 796
24 more
  1. 2019Learning Representations by Maximizing Mutual Information Across Viewscited by 733
  2. 2017Poincar'e Embeddings for Learning Hierarchical Representationscited by 674
  3. 2019Data-Efficient Image Recognition with Contrastive Predictive Codingcited by 672
  4. 2018Hyperbolic Neural Networkscited by 393
  5. 2015What's the Point: Semantic Segmentation with Point Supervisioncited by 297
  6. 2016Learning Representations for Automatic Colorizationcited by 251
  7. 2015BoxSup: Exploiting Bounding Boxes to Supervise Convolutional Networks for Semantic Segmentationcited by 238
  8. 2016Simple Does It: Weakly Supervised Instance and Semantic Segmentationcited by 226
  9. 2019Hyperbolic Image Embeddingscited by 219
  10. 2017Colorization as a Proxy Task for Visual Understandingcited by 175
  11. 2014Simultaneous Detection and Segmentationcited by 158
  12. 2017Semantic Instance Segmentation with a Discriminative Loss Functioncited by 118
  13. 2019Continuous Hierarchical Representations with Poincar'e Variational Auto-Encoderscited by 111
  14. 2019SCOPS: Self-Supervised Co-Part Segmentationcited by 100
  15. 2018Weakly Supervised Instance Segmentation using Class Peak Responsecited by 74
  16. 2015Unsupervised Object Discovery and Localization in the Wild: Part-based Matching with Bottom-up Region Proposalscited by 69
  17. 2017Learning to Segment Every Thingcited by 64
  18. 2017Cross-Domain Self-supervised Multi-task Feature Learning using Synthetic Imagerycited by 56
  19. 2017MaskLab: Instance Segmentation by Refining Object Detection with Semantic and Direction Featurescited by 54
  20. 2019ShapeMask: Learning to Segment Novel Objects by Refining Shape Priorscited by 42
  21. 2017Semantic Instance Segmentation via Deep Metric Learningcited by 37
  22. 2017Improvements to context based self-supervised learningcited by 32
  23. 2018Learning Instance Segmentation by Interactioncited by 18
  24. 2020Commonality-Parsing Network across Shape and Appearance for Partially Supervised Instance Segmentationcited by 14

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 3; 2022: 9; 2023: 7; 2024: 11; 2025: 8; 2026: 1.

39 papers.

Sort:

  1. 2021Deep Long-Tailed Learning: A Surveycited by 416
  2. 2022Hyperbolic Image Segmentationcited by 101
  3. 2021Self-Supervised Aggregation of Diverse Experts for Test-Agnostic Long-Tailed Recognitioncited by 95
  4. 2023Hyperbolic Deep Learning in Computer Vision: A Surveycited by 72
  5. 2022Hyperbolic Contrastive Learning for Visual Representations beyond Objectscited by 58
  6. 2021CO-SNE: Dimensionality Reduction and Visualization for Hyperbolic Datacited by 28
  7. 2024A Systematic Review on Long-Tailed Learningcited by 27
  8. 2022ProCo: Prototype-aware Contrastive Learning for Long-tailed Medical Image Classificationcited by 25
  9. 2023Poincar'e ResNetcited by 24
  10. 2024Hyperbolic Fine-Tuning for Large Language Modelscited by 24
  11. 2025Hyperbolic Category Discoverycited by 20
  12. 2023Novel Class Discovery for Long-tailed Recognitioncited by 19
27 more
  1. 2022Rethinking the compositionality of point clouds through regularization in the hyperbolic spacecited by 18
  2. 2022Towards Calibrated Hyper-Sphere Representation via Distribution Overlap Coefficient for Long-tailed Learningcited by 17
  3. 2024Lorentzian Residual Neural Networkscited by 17
  4. 2023Hyperbolic Contrastive Learningcited by 15
  5. 2025HyperCore: The Core Framework for Building Hyperbolic Foundation Models with Comprehensive Modulescited by 12
  6. 2022Learning to Discover and Detect Objectscited by 11
  7. 2024Improving Visual Recognition with Hyperbolical Visual Hierarchy Mappingcited by 11
  8. 2024Understanding Hyperbolic Metric Learning through Hard Negative Samplingcited by 10
  9. 2022HP-Capsule: Unsupervised Face Part Discovery by Hierarchical Parsing Capsule Networkcited by 7
  10. 2024Lidar Panoptic Segmentation in an Open Worldcited by 7
  11. 2023Hyperbolic Geometry in Computer Vision: A Surveycited by 4
  12. 2024Hyp-OC: Hyperbolic One Class Classification for Face Anti-Spoofingcited by 3
  13. 2024Adversarial Attacks on Hyperbolic Networkscited by 3
  14. 2023Unsupervised Feature Learning with Emergent Data-Driven Prototypicalitycited by 2
  15. 2023Enhancing Novel Object Detection via Cooperative Foundational Modelscited by 2
  16. 2024PANDAS: Prototype-based Novel Class Discovery and Detectioncited by 2
  17. 2024Open-World Panoptic Segmentationcited by 2
  18. 2025Generalized Class Discovery in Instance Segmentationcited by 2
  19. 2025Image Segmentation with Large Language Models: A Survey with Perspectives for Intelligent Transportation Systemscited by 2
  20. 2025HIDISC: A Hyperbolic Framework for Domain Generalization with Generalized Category Discoverycited by 2
  21. 2022CYBORGS: Contrastively Bootstrapping Object Representations by Grounding in Segmentationcited by 1
  22. 2022Teacher Guided Training: An Efficient Framework for Knowledge Transfercited by 1
  23. 2025Predicate Hierarchies Improve Few-Shot State Classificationcited by 1
  24. 2025LIHE: Linguistic Instance-Split Hyperbolic-Euclidean Framework for Generalized Weakly-Supervised Referring Expression Comprehensioncited by 1
  25. 2024A Bottom-Up Approach to Class-Agnostic Image Segmentationcited by 0
  26. 2025HypeVPR: Exploring Hyperbolic Space for Perspective to Equirectangular Visual Place Recognitioncited by 0
  27. 2026Lorentz Framework for Semantic Segmentationcited by 0