mlsanityData to 14 Oct 2026

SegDA: Maximum Separable Segment Mask with Pseudo Labels for Domain Adaptive Semantic Segmentation

arXiv 2308.05851 ↗First posted 10 Aug 2023cs.CVcited by 2

Not on the map. Why?

Abstract (arXiv)

Unsupervised Domain Adaptation (UDA) aims to solve the problem of label scarcity of the target domain by transferring the knowledge from the label rich source domain. Usually, the source domain consists of synthetic images for which the annotation is easily obtained using the well known computer graphics techniques. However, obtaining annotation for real world images (target domain) require lot of manual annotation effort and is very time consuming because it requires per pixel annotation. To address this problem we propose SegDA module to enhance transfer performance of UDA methods by learning the maximum separable segment representation. This resolves the problem of identifying visually similar classes like pedestrian/rider, sidewalk/road etc. We leveraged Equiangular Tight Frame (ETF) classifier inspired from Neural Collapse for maximal separation between segment classes. This causes the source domain pixel representation to collapse to a single vector forming a simplex vertices which are aligned to the maximal separable ETF classifier. We use this phenomenon to propose the novel architecture for domain adaptation of segment representation for target domain. Additionally, we proposed to estimate the noise in labelling the target domain images and update the decoder for noise correction which encourages the discovery of pixels for classes not identified in pseudo labels. We have used four UDA benchmarks simulating synthetic-to-real, daytime-to-nighttime, clear-to-adverse weather scenarios. Our proposed approach outperforms +2.2 mIoU on GTA -> Cityscapes, +2.0 mIoU on Synthia -> Cityscapes, +5.9 mIoU on Cityscapes -> DarkZurich, +2.6 mIoU on Cityscapes -> ACDC.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 55 of its 69 references that are in the map. First posted on arXiv 10 Aug 2023.

Builds on

55 of 69 references.

References · 55

  1. 2017Attention Is All You Needcited by 60,361
  2. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  3. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  4. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,139
  5. 2020End-to-End Object Detection with Transformerscited by 7,414
  6. 2014Fully Convolutional Networks for Semantic Segmentationcited by 4,859
  7. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,444
  8. 2016DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFscited by 3,241
  9. 2021SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformerscited by 2,940
  10. 2020FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidencecited by 2,085
  11. 2018Unsupervised Representation Learning by Predicting Image Rotationscited by 1,591
  12. 2020Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspective with Transformerscited by 1,369
38 more
  1. 2017Random Erasing Data Augmentationcited by 1,239
  2. 2021Per-Pixel Classification is Not All You Need for Semantic Segmentationcited by 1,152
  3. 2016Colorful Image Colorizationcited by 1,114
  4. 2017CyCADA: Cycle-Consistent Adversarial Domain Adaptationcited by 1,082
  5. 2016Playing for Data: Ground Truth from Computer Gamescited by 818
  6. 2020Prevalence of Neural Collapse during the terminal phase of deep learning trainingcited by 719
  7. 2017Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networkscited by 707
  8. 2018ADVENT: Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentationcited by 623
  9. 2018Learning to Adapt Structured Output Space for Semantic Segmentationcited by 584
  10. 2020FDA: Fourier Domain Adaptation for Semantic Segmentationcited by 541
  11. 2020Dense Contrastive Learning for Self-Supervised Visual Pre-Trainingcited by 512
  12. 2019Confidence Regularized Self-Trainingcited by 419
  13. 2019Bidirectional Learning for Domain Adaptation of Semantic Segmentationcited by 308
  14. 2021DAFormer: Improving Network Architectures and Training Strategies for Domain-Adaptive Semantic Segmentationcited by 300
  15. 2018Taking A Closer Look at Domain Shift: Category-level Adversaries for Semantics Consistent Domain Adaptationcited by 295
  16. 2021Prototypical Pseudo Label Denoising and Target Structure Learning for Domain Adaptive Semantic Segmentationcited by 268
  17. 2020DACS: Domain Adaptation via Cross-domain Mixed Samplingcited by 226
  18. 2020Rectifying Pseudo Label Learning via Uncertainty Estimation for Domain Adaptive Semantic Segmentationcited by 222
  19. 2020Unsupervised Intra-domain Adaptation for Semantic Segmentation through Self-Supervisioncited by 198
  20. 2021Unsupervised Semantic Segmentation by Contrasting Object Mask Proposalscited by 190
  21. 2020Instance Adaptive Self-Training for Unsupervised Domain Adaptationcited by 177
  22. 2022HRDA: Context-Aware High-Resolution Domain-Adaptive Semantic Segmentationcited by 164
  23. 2019Domain Adaptation for Structured Output via Discriminative Patch Representationscited by 150
  24. 2021Self-supervised Augmentation Consistency for Adapting Semantic Segmentationcited by 149
  25. 2019Domain Adaptation for Semantic Segmentation with Maximum Squares Losscited by 138
  26. 2018DCAN: Dual Channel-wise Alignment Networks for Unsupervised Scene Adaptationcited by 101
  27. 2022SePiCo: Semantic-Guided Pixel Contrast for Domain Adaptive Semantic Segmentationcited by 87
  28. 2020Map-Guided Curriculum Domain Adaptation and Uncertainty-Aware Evaluation for Semantic Nighttime Image Segmentationcited by 86
  29. 2021DANNet: A One-Stage Domain Adaptation Network for Unsupervised Nighttime Semantic Segmentationcited by 80
  30. 2021Context-Aware Mixup for Domain Adaptive Semantic Segmentationcited by 79
  31. 2019Significance-aware Information Bottleneck for Domain Adaptive Semantic Segmentationcited by 74
  32. 2020Three Ways to Improve Semantic Segmentation with Self-Supervised Depth Estimationcited by 66
  33. 2020Pixel-Level Cycle Association: A New Perspective for Domain Adaptive Semantic Segmentationcited by 58
  34. 2019Unsupervised Scene Adaptation with Memory Regularization in vivocited by 57
  35. 2020Uncertainty-Aware Consistency Regularization for Cross-Domain Semantic Segmentationcited by 45
  36. 2019ACE: Adapting to Changing Environments for Semantic Segmentationcited by 44
  37. 2023Neural Collapse in Deep Linear Networks: From Balanced to Imbalanced Datacited by 40
  38. 2022Multiple-environment Self-adaptive Network for Aerial-view Geo-localizationcited by 37

50 most cited of 55 All 55 (JSON)

Built on by

Cited by year

Citing papers per year: 2023: 1; 2024: 1.

2 papers.

  1. 2023FloCoDe: Unbiased Dynamic Scene Graph Generation with Temporal Consistency and Correlation Debiasingcited by 6
  2. 2024Layer-wise Model Merging for Unsupervised Domain Adaptation in Segmentation Taskscited by 1