mlsanityData to 14 Oct 2026

Incremental Learning in Semantic Segmentation from Image Labels

arXiv 2112.01882 ↗First posted 3 Dec 2021cs.CVcited by 39

Not on the map. Why?

Abstract (arXiv)

Although existing semantic segmentation approaches achieve impressive results, they still struggle to update their models incrementally as new categories are uncovered. Furthermore, pixel-by-pixel annotations are expensive and time-consuming. This paper proposes a novel framework for Weakly Incremental Learning for Semantic Segmentation, that aims at learning to segment new classes from cheap and largely available image-level labels. As opposed to existing approaches, that need to generate pseudo-labels offline, we use an auxiliary classifier, trained with image-level labels and regularized by the segmentation model, to obtain pseudo-supervision online and update the model incrementally. We cope with the inherent noise in the process by using soft-labels generated by the auxiliary classifier. We demonstrate the effectiveness of our approach on the Pascal VOC and COCO datasets, outperforming offline weakly-supervised methods and obtaining results comparable with incremental learning methods with full supervision. Code can be found at https://github.com/fcdl94/WILSON.

Key facts. Cited by 39 papers in the map (data through 14 Oct 2026). Builds on 59 of its 70 references that are in the map. First posted on arXiv 3 Dec 2021.

Builds on

59 of 70 references.

References · 59

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2014Microsoft COCO: Common Objects in Contextcited by 17,608
  3. 2015Distilling the Knowledge in a Neural Networkcited by 10,324
  4. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,444
  5. 2016DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFscited by 3,241
  6. 2018Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentationcited by 3,223
  7. 2016Learning without Forgettingcited by 2,496
  8. 2015Learning Deep Features for Discriminative Localizationcited by 2,436
  9. 2016iCaRL: Incremental Classifier and Representation Learningcited by 2,333
  10. 2017Rethinking Atrous Convolution for Semantic Image Segmentationcited by 2,161
  11. 2019Energy and Policy Considerations for Deep Learning in NLPcited by 1,254
  12. 2017Continual Learning with Deep Generative Replaycited by 1,158
38 more
  1. 2016COCO-Stuff: Thing and Stuff Classes in Contextcited by 864
  2. 2016Playing for Data: Ground Truth from Computer Gamescited by 818
  3. 2018Riemannian Walk for Incremental Learning: Understanding Forgetting and Intransigencecited by 742
  4. 2021Carbon Emissions and Large Neural Network Trainingcited by 567
  5. 2016ScribbleSup: Scribble-Supervised Convolutional Networks for Semantic Segmentationcited by 331
  6. 2018Learning Pixel-level Semantic Affinity with Image-level Supervision for Weakly Supervised Semantic Segmentationcited by 321
  7. 2020Self-supervised Equivariant Attention Mechanism for Weakly Supervised Semantic Segmentationcited by 300
  8. 2015What's the Point: Semantic Segmentation with Point Supervisioncited by 298
  9. 2018Learning without Memorizingcited by 250
  10. 2019Weakly Supervised Learning of Instance Segmentation with Inter-pixel Relationscited by 247
  11. 2015BoxSup: Exploiting Bounding Boxes to Supervise Convolutional Networks for Semantic Segmentationcited by 238
  12. 2016Simple Does It: Weakly Supervised Instance and Semantic Segmentationcited by 227
  13. 2020Mnemonics Training: Multi-Class Incremental Learning without Forgettingcited by 221
  14. 2020Does label smoothing mitigate label noise?cited by 204
  15. 2017Object Region Mining with Adversarial Erasing: A Simple Classification to Semantic Segmentation Approachcited by 201
  16. 2016Seed, Expand and Constrain: Three Principles for Weakly-Supervised Image Segmentationcited by 198
  17. 2020The Computational Limits of Deep Learningcited by 193
  18. 2020Modeling the Background for Incremental Learning in Semantic Segmentationcited by 190
  19. 2015Weakly- and Semi-Supervised Learning of a DCNN for Semantic Image Segmentationcited by 189
  20. 2016Wider or Deeper: Revisiting the ResNet Model for Visual Recognitioncited by 184
  21. 2020PLOP: Learning without Forgetting for Continual Semantic Segmentationcited by 177
  22. 2020Single-Stage Semantic Segmentation from Image Labelscited by 167
  23. 2019FickleNet: Weakly and Semi-supervised Semantic Image Segmentation using Stochastic Inferencecited by 162
  24. 2020Weakly-Supervised Semantic Segmentation via Sub-category Explorationcited by 145
  25. 2020Mining Cross-Image Semantics for Weakly Supervised Semantic Segmentationcited by 144
  26. 2019Incremental Learning Techniques for Semantic Segmentationcited by 141
  27. 2021Railroad is not a Train: Saliency as Pseudo-pixel Supervision for Weakly Supervised Semantic Segmentationcited by 127
  28. 2020SS-IL: Separated Softmax for Incremental Learningcited by 126
  29. 2021Distilling Causal Effect of Data in Class-Incremental Learningcited by 117
  30. 2018Normalized Cut Loss for Weakly-supervised CNN Segmentationcited by 112
  31. 2018Learning random-walk label propagation for weakly-supervised semantic segmentationcited by 101
  32. 2015Constrained Convolutional Neural Networks for Weakly Supervised Segmentationcited by 100
  33. 2021Continual Semantic Segmentation via Repulsion-Attraction of Sparse and Disentangled Latent Representationscited by 96
  34. 2018Self-Erasing Network for Integral Object Attentioncited by 95
  35. 2021Non-Salient Region Object Mining for Weakly Supervised Semantic Segmentationcited by 95
  36. 2021Pointly-Supervised Instance Segmentationcited by 83
  37. 2021RECALL: Replay-based Continual Learning in Semantic Segmentationcited by 83
  38. 2017In-Place Activated BatchNorm for Memory-Optimized Training of DNNscited by 81

50 most cited of 59 All 59 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 6; 2023: 17; 2024: 8; 2025: 6; 2026: 2.

39 papers.

Sort:

By area:

  1. 2023A Comprehensive Survey of Continual Learning: Theory, Method and Applicationcited by 872
  2. 2023Class-Incremental Learning: A Surveycited by 209
  3. 2023Revisiting Class-Incremental Learning with Pre-Trained Models: Generalizability and Adaptivity are All You Needcited by 184
  4. 2023Unmasking Anomalies in Road-Scene Segmentationcited by 45
  5. 2023Audio-Visual Class-Incremental Learningcited by 33
  6. 2023Semantic Image Segmentation: Two Decades of Researchcited by 27
  7. 2022CoMFormer: Continual Learning in Semantic and Panoptic Segmentationcited by 25
  8. 2022Modeling Missing Annotations for Incremental Learning in Object Detectioncited by 23
  9. 2023A Survey on Continual Semantic Segmentation: Theory, Challenge, Method and Applicationcited by 21
  10. 2023Mask2Anomaly: Mask Transformer for Universal Open-set Segmentationcited by 18
  11. 2023Foundation Model Drives Weakly Incremental Learning for Semantic Segmentationcited by 12
  12. 2022Learning with Style: Continual Semantic Segmentation Across Tasks and Domainscited by 9
27 more
  1. 2023Continual Learning for LiDAR Semantic Segmentation: Class-Incremental and Coarse-to-Fine strategies on Sparse Datacited by 9
  2. 2024Continual Segmentation with Disentangled Objectness Learning and Class Recognitioncited by 8
  3. 2022Online Unsupervised Domain Adaptation for Person Re-identificationcited by 7
  4. 2023RaSP: Relation-aware Semantic Prior for Weakly Supervised Incremental Segmentationcited by 6
  5. 2023Continual Road-Scene Semantic Segmentation via Feature-Aligned Symmetric Multi-Modal Networkcited by 6
  6. 2025CoMBO: Conflict Mitigation via Branched Optimization for Class Incremental Segmentationcited by 6
  7. 2024Learning from the Web: Language Drives Weakly-Supervised Incremental Learning for Semantic Segmentationcited by 5
  8. 2022Causes of Catastrophic Forgetting in Class-Incremental Semantic Segmentationcited by 4
  9. 2023Principles of Forgetting in Domain-Incremental Semantic Segmentation in Adverse Weather Conditionscited by 4
  10. 2023RECALL+: Adversarial Web-based Replay for Continual Learning in Semantic Segmentationcited by 4
  11. 2024Scribbles for All: Benchmarking Scribble Supervised Segmentation Across Datasetscited by 4
  12. 2023Effects of Architectures on Continual Semantic Segmentationcited by 3
  13. 2024Source-Guided Similarity Preservation for Online Person Re-Identificationcited by 3
  14. 2024Tendency-driven Mutual Exclusivity for Weakly Supervised Incremental Semantic Segmentationcited by 3
  15. 2023Towards Label-Efficient Incremental Learning: A Surveycited by 2
  16. 2024PANDAS: Prototype-based Novel Class Discovery and Detectioncited by 2
  17. 2024Pixel-Level Domain Adaptation: A New Perspective for Enhancing Weakly Supervised Semantic Segmentationcited by 2
  18. 2025Continual Multiple Instance Learning with Enhanced Localization for Histopathological Whole Slide Image Analysiscited by 2
  19. 2022Delving into Transformer for Incremental Semantic Segmentationcited by 1
  20. 2025CLIMB-3D: Continual Learning for Imbalanced 3D Instance Segmentationcited by 1
  21. 2025Completely Weakly Supervised Class-Incremental Learning for Semantic Segmentationcited by 1
  22. 2025Learning Yourself: Class-Incremental Semantic Segmentation with Language-Inspired Bootstrapped Disentanglementcited by 1
  23. 2026ACIL: Active Class Incremental Learning for Image Classificationcited by 1
  24. 2023Vision-Based Cranberry Crop Ripening Assessmentcited by 0
  25. 2024Incremental Label Distribution Learning with Scalable Graph Convolutional Networkscited by 0
  26. 2025Incremental Human-Object Interaction Detection with Invariant Relation Representation Learningcited by 0
  27. 2026DR.WILSS: Diffusion-Based Replay for Weakly Supervised Continual Semantic Segmentationcited by 0