mlsanityData to 7 Oct 2026

Generalize then Adapt: Source-Free Domain Adaptive Semantic Segmentation

arXiv 2108.11249 ↗First posted 25 Aug 2021cs.CV cs.LGcited by 81

Not on the map. Why?

Abstract (arXiv)

Unsupervised domain adaptation (DA) has gained substantial interest in semantic segmentation. However, almost all prior arts assume concurrent access to both labeled source and unlabeled target, making them unsuitable for scenarios demanding source-free adaptation. In this work, we enable source-free DA by partitioning the task into two: a) source-only domain generalization and b) source-free target adaptation. Towards the former, we provide theoretical insights to develop a multi-head framework trained with a virtually extended multi-source dataset, aiming to balance generalization and specificity. Towards the latter, we utilize the multi-head framework to extract reliable target pseudo-labels for self-training. Additionally, we introduce a novel conditional prior-enforcing auto-encoder that discourages spatial irregularities, thereby enhancing the pseudo-label quality. Experiments on the standard GTA5-to-Cityscapes and SYNTHIA-to-Cityscapes benchmarks show our superiority even against the non-source-free prior-arts. Further, we show our compatibility with online adaptation enabling deployment in a sequentially changing environment.

Key facts. Cited by 81 papers in the map (data through 7 Oct 2026). Builds on 83 of its 102 references that are in the map. First posted on arXiv 25 Aug 2021.

Builds on

83 of 102 references.

References · 83

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  4. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
  5. 2014Fully Convolutional Networks for Semantic Segmentationcited by 4,845
  6. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,427
  7. 2016DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFscited by 3,239
  8. 2017Arbitrary Style Transfer in Real-time with Adaptive Instance Normalizationcited by 1,998
  9. 2017CyCADA: Cycle-Consistent Adversarial Domain Adaptationcited by 1,079
  10. 2017Deeper, Broader and Artier Domain Generalizationcited by 983
  11. 2020Do We Really Need to Access the Source Data? Source Hypothesis Transfer for Unsupervised Domain Adaptationcited by 844
  12. 2016Playing for Data: Ground Truth from Computer Gamescited by 816
38 more
  1. 2017Maximum Classifier Discrepancy for Unsupervised Domain Adaptationcited by 682
  2. 2018ADVENT: Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentationcited by 623
  3. 2018Learning to Adapt Structured Output Space for Semantic Segmentationcited by 584
  4. 2020FDA: Fourier Domain Adaptation for Semantic Segmentationcited by 540
  5. 2019Confidence Regularized Self-Trainingcited by 418
  6. 2018Two at Once: Enhancing Learning and Generalization Capacities via IBN-Netcited by 361
  7. 2019Benchmarking Robustness in Object Detection: Autonomous Driving when Winter is Comingcited by 332
  8. 2019Bidirectional Learning for Domain Adaptation of Semantic Segmentationcited by 308
  9. 2018Taking A Closer Look at Domain Shift: Category-level Adversaries for Semantics Consistent Domain Adaptationcited by 295
  10. 2019Domain Randomization and Pyramid Consistency: Simulation-to-Real Generalization without Accessing Target Domain Datacited by 245
  11. 2016FCNs in the Wild: Pixel-level Adversarial and Constraint-based Adaptationcited by 242
  12. 2020DACS: Domain Adaptation via Cross-domain Mixed Samplingcited by 226
  13. 2020Rectifying Pseudo Label Learning via Uncertainty Estimation for Domain Adaptive Semantic Segmentationcited by 222
  14. 2020Universal Source-Free Domain Adaptationcited by 219
  15. 2020Unsupervised Intra-domain Adaptation for Semantic Segmentation through Self-Supervisioncited by 198
  16. 2020Instance Adaptive Self-Training for Unsupervised Domain Adaptationcited by 177
  17. 2018DLOW: Domain Flow for Adaptation and Generalizationcited by 176
  18. 2021Source-Free Domain Adaptation for Semantic Segmentationcited by 168
  19. 2020Learning Texture Invariant Representation for Domain Adaptation of Semantic Segmentationcited by 153
  20. 2019Category Anchor-Guided Unsupervised Domain Adaptation for Semantic Segmentationcited by 151
  21. 2019Domain Adaptation for Structured Output via Discriminative Patch Representationscited by 150
  22. 2020Classes Matter: A Fine-grained Adversarial Approach to Cross-domain Semantic Segmentationcited by 142
  23. 2019Domain Adaptation for Semantic Segmentation with Maximum Squares Losscited by 138
  24. 2020Efficient Domain Generalization via Common-Specific Low-Rank Decompositioncited by 134
  25. 2019Self-Ensembling with GAN-based Data Augmentation for Domain Adaptation in Semantic Segmentationcited by 126
  26. 2020Differential Treatment for Stuff and Things: A Simple Unsupervised Domain Adaptation Method for Semantic Segmentationcited by 126
  27. 2018Learning Semantic Segmentation from Synthetic Data: A Geometrically Guided Input-Output Adaptation Approachcited by 119
  28. 2017No More Discrimination: Cross City Adaptation of Road Scene Segmenterscited by 110
  29. 2019Domain Agnostic Learning with Disentangled Representationscited by 107
  30. 2020CrDoCo: Pixel-level Domain Transfer with Cross-Domain Consistencycited by 107
  31. 2019All about Structure: Adapting Structural Information across Domains for Boosting Semantic Segmentationcited by 104
  32. 2019Constructing Self-motivated Pyramid Curriculums for Cross-Domain Semantic Segmentation: A Non-Adversarial Approachcited by 103
  33. 2018DCAN: Dual Channel-wise Alignment Networks for Unsupervised Scene Adaptationcited by 101
  34. 2018Model Adaptation with Synthetic and Real Data for Semantic Dense Foggy Scene Understandingcited by 98
  35. 2019DADA: Depth-aware Domain Adaptation in Semantic Segmentationcited by 97
  36. 2017ROAD: Reality Oriented Adaptation for Semantic Segmentation of Urban Scenescited by 95
  37. 2016Fast Patch-based Style Transfer of Arbitrary Stylecited by 92
  38. 2020Unsupervised Domain Adaptation in Semantic Segmentation: a Reviewcited by 88

50 most cited of 83 All 83 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 6; 2022: 21; 2023: 30; 2024: 12; 2025: 9; 2026: 3.

81 papers.

Sort:

By area:

  1. 2022Continual Test-Time Domain Adaptationcited by 520
  2. 2023A Comprehensive Survey on Test-Time Adaptation under Distribution Shiftscited by 353
  3. 2022A Closer Look at Smoothness in Domain Adversarial Trainingcited by 85
  4. 2022Source-Free Unsupervised Domain Adaptation: A Surveycited by 79
  5. 2022Balancing Discriminability and Transferability for Source-Free Domain Adaptationcited by 58
  6. 2021Towards Fewer Annotations: Active Learning via Region Impurity and Prediction Uncertainty for Domain Adaptive Semantic Segmentationcited by 56
  7. 2022Instance Relation Graph Guided Source-Free Domain Adaptive Object Detectioncited by 55
  8. 2022PASTA: Proportional Amplitude Spectrum Training Augmentation for Syn-to-Real Domain Generalizationcited by 38
  9. 2022Towards Online Domain Adaptive Object Detectioncited by 36
  10. 2022Learning Across Domains and Devices: Style-Driven Source-Free Domain Adaptation in Clustered Federated Learningcited by 35
  11. 2024Semantics, Distortion, and Style Matter: Towards Source-free UDA for Panoramic Segmentationcited by 34
  12. 2021Data-Free Knowledge Transfer: A Surveycited by 31
38 more
  1. 2023Improving the Generalization of Segmentation Foundation Model under Distribution Shift via Weakly Supervised Adaptationcited by 31
  2. 2021Generation, augmentation, and alignment: A pseudo-source domain based method for source-free domain adaptationcited by 28
  3. 2024360SFUDA++: Towards Source-free UDA for Panoramic Segmentation by Learning Reliable Category Prototypescited by 28
  4. 2023Semantic Image Segmentation: Two Decades of Researchcited by 26
  5. 2022Concurrent Subsidiary Supervision for Unsupervised Source-Free Domain Adaptationcited by 21
  6. 2022Ground then Navigate: Language-guided Navigation in Dynamic Scenescited by 20
  7. 2023Multi-Modal Continual Test-Time Adaptation for 3D Semantic Segmentationcited by 20
  8. 2021Unsupervised Domain Adaptation for Semantic Image Segmentation: a Comprehensive Surveycited by 19
  9. 2023DiGA: Distil to Generalize and then Adapt for Domain Adaptive Semantic Segmentationcited by 19
  10. 2023Hyperbolic Active Learning for Semantic Segmentation under Domain Shiftcited by 18
  11. 2022SimT: Handling Open-set Noise for Domain Adaptive Semantic Segmentationcited by 17
  12. 2022Subsidiary Prototype Alignment for Universal Domain Adaptationcited by 17
  13. 2023Domain-Specificity Inducing Transformers for Source-Free Domain Adaptationcited by 16
  14. 2022Memory Consistent Unsupervised Off-the-Shelf Model Adaptation for Source-Relaxed Medical Image Segmentationcited by 15
  15. 2024Training-Free Model Merging for Multi-target Domain Adaptationcited by 15
  16. 2022Unsupervised Model Adaptation for Source-free Segmentation of Medical Imagescited by 13
  17. 2022Continual BatchNorm Adaptation (CBNA) for Semantic Segmentationcited by 12
  18. 2023Contrastive Model Adaptation for Cross-Condition Robustness in Semantic Segmentationcited by 12
  19. 2023Spatio-Temporal Pixel-Level Contrastive Learning-based Source-Free Domain Adaptation for Video Semantic Segmentationcited by 10
  20. 2023Cal-SFDA: Source-Free Domain-adaptive Semantic Segmentation with Differentiable Expected Calibration Errorcited by 10
  21. 2022Anatomy-guided domain adaptation for 3D in-bed human pose estimationcited by 9
  22. 2023Towards Source-free Domain Adaptive Semantic Segmentation via Importance-aware and Prototype-contrast Learningcited by 9
  23. 2022Semantic Self-adaptation: Enhancing Generalization with a Single Samplecited by 8
  24. 2023Generative appearance replay for continual unsupervised domain adaptationcited by 8
  25. 2023Source-Free Domain Adaptation for RGB-D Semantic Segmentation with Vision Transformerscited by 8
  26. 2023Learning Class and Domain Augmentations for Single-Source Open-Domain Generalizationcited by 8
  27. 2024Stable Neighbor Denoising for Source-free Domain Adaptive Segmentationcited by 8
  28. 2022Amplitude Spectrum Transformation for Open Compound Domain Adaptive Semantic Segmentationcited by 7
  29. 2023Iterative Loop Method Combining Active and Semi-Supervised Learning for Domain Adaptive Semantic Segmentationcited by 7
  30. 2023Pulling Target to Source: A New Perspective on Domain Adaptive Semantic Segmentationcited by 7
  31. 2025CoPT: Unsupervised Domain Adaptive Segmentation using Domain-Agnostic Text Embeddingscited by 7
  32. 2021Unsupervised Adaptation of Semantic Segmentation Models without Source Datacited by 6
  33. 2024ControlUDA: Controllable Diffusion-assisted Unsupervised Domain Adaptation for Cross-Weather Semantic Segmentationcited by 6
  34. 2024Train Till You Drop: Towards Stable and Robust Source-free Unsupervised 3D Domain Adaptationcited by 6
  35. 2023Aligning Non-Causal Factors for Transformer-Based Source-Free Domain Adaptationcited by 5
  36. 2024GLC++: Source-Free Universal Domain Adaptation through Global-Local Clustering and Contrastive Affinity Learningcited by 4
  37. 2024Efficient Source-Free Time-Series Adaptation via Parameter Subspace Disentanglementcited by 4
  38. 2025BatStyler: Advancing Multi-category Style Generation for Source-free Domain Generalizationcited by 4

50 most cited of 81 All 81 (JSON)