Feature Re-Representation and Reliable Pseudo Label Retraining for Cross-Domain Semantic Segmentation

This paper presents a novel unsupervised domain adaptation method for semantic segmentation. We argue that a good representation of the target-domain data should keep both the knowledge from the source domain and the target-domain-specific information. To obtain the knowledge from the source domain, we first learn a set of bases to characterize the feature distribution of the source domain, then features from both the source and the target domain are re-represented as a weighted summation of the source bases. A discriminator is additionally introduced to make the re-representation responsibilities of both domain features under the same bases indistinguishable. In this way, the domain gap between the source re-representation and target re-representation is minimized, and the re-represented target domain features contain the source domain information. Then we combine the feature re-representation with the original domain-specific feature together for subsequent pixel-wise classification. To further make the re-represented target features semantically meaningful, a Reliable Pseudo Label Retraining (RPLR) strategy is proposed, which utilizes the consistency of the prediction by the networks trained with multi-view source images to select the clean pseudo labels on unlabeled target images for re-training. Extensive experiments demonstrate the competitive performance of our approach for unsupervised domain adaptation on the semantic segmentation benchmarks.

• Publication year

  2022

• Venue

  IEEE Transactions on Pattern Analysis and Machine Intelligence

• Publication date

  2022-03-11

• Fields of study

  Medicine, Computer Science

• Identifiers
  DOI 10.1109/TPAMI.2022.3154933PMID 35275810
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar, PubMed

• No claims are published for this paper.

• No concepts are published for this paper.

• Differential Treatment for Stuff and Things: A Simple Unsupervised Domain Adaptation Method for Semantic Segmentation

  2020influential reference
• Unsupervised Intra-Domain Adaptation for Semantic Segmentation Through Self-Supervision

  2020influential reference
• Domain Adaptive Semantic Segmentation Using Weak Labels

  2020cited by this paper
• Learning Texture Invariant Representation for Domain Adaptation of Semantic Segmentation

  2020cited by this paper
• Contextual-Relation Consistent Domain Adaptation for Semantic Segmentation

  2020cited by this paper
• AUTO-ENCODING VARIATIONAL BAYES

  2020cited by this paper
• FDA: Fourier Domain Adaptation for Semantic Segmentation

  2020cited by this paper
• Self-Ensembling With GAN-Based Data Augmentation for Domain Adaptation in Semantic Segmentation

  2019cited by this paper
• Confidence Regularized Self-Training

  2019influential reference
• Expectation-Maximization Attention Networks for Semantic Segmentation

  2019influential reference
• Bidirectional Learning for Domain Adaptation of Semantic Segmentation

  2019influential reference
• Transferable Adversarial Training: A General Approach to Adapting Deep Classifiers

  2019cited by this paper
• All About Structure: Adapting Structural Information Across Domains for Boosting Semantic Segmentation

  2019cited by this paper
• Self-Ensembling Attention Networks: Addressing Domain Shift for Semantic Segmentation

  2019cited by this paper
• Domain Adaptation for Structured Output via Discriminative Patch Representations

  2019cited by this paper
• SSF-DAN: Separated Semantic Feature Based Domain Adaptation Network for Semantic Segmentation

  2019cited by this paper
• Addressing Failure Prediction by Learning Model Confidence

  2019cited by this paper
• Domain Adaptation for Semantic Segmentation With Maximum Squares Loss

  2019cited by this paper
• ExFuse: Enhancing Feature Fusion for Semantic Segmentation

  2018cited by this paper
• Learning to Adapt Structured Output Space for Semantic Segmentation

  2018influential reference
• Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation

  2018cited by this paper
• DCAN: Dual Channel-wise Alignment Networks for Unsupervised Scene Adaptation

  2018cited by this paper
• To Trust Or Not To Trust A Classifier

  2018cited by this paper
• Dual Attention Network for Scene Segmentation

  2018cited by this paper
• Taking a Closer Look at Domain Shift: Category-Level Adversaries for Semantics Consistent Domain Adaptation

  2018cited by this paper
• Unsupervised Domain Adaptation for Semantic Segmentation via Class-Balanced Self-training

  2018influential reference
• Domain Transfer Through Deep Activation Matching

  2018cited by this paper
• PSANet: Point-wise Spatial Attention Network for Scene Parsing

  2018cited by this paper
• CCNet: Criss-Cross Attention for Semantic Segmentation

  2018cited by this paper
• ADVENT: Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentation

  2018cited by this paper
• DLOW: Domain Flow for Adaptation and Generalization

  2018cited by this paper
• ROAD: Reality Oriented Adaptation for Semantic Segmentation of Urban Scenes

  2017cited by this paper
• Asymmetric Tri-training for Unsupervised Domain Adaptation

  2017cited by this paper
• A Fully Convolutional Tri-Branch Network (FCTN) for Domain Adaptation

  2017cited by this paper
• Fully Convolutional Networks for Semantic Segmentation

  2017cited by this paper
• Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results

  2017cited by this paper
• Rethinking Atrous Convolution for Semantic Image Segmentation

  2017cited by this paper
• Vector Quantization

  2017cited by this paper
• Large Kernel Matters — Improve Semantic Segmentation by Global Convolutional Network

  2017cited by this paper
• Image to Image Translation for Domain Adaptation

  2017cited by this paper
• CyCADA: Cycle-Consistent Adversarial Domain Adaptation

  2017influential reference
• Attention is All you Need

  2017cited by this paper
• Non-local Neural Networks

  2017cited by this paper
• The SYNTHIA Dataset: A Large Collection of Synthetic Images for Semantic Segmentation of Urban Scenes

  2016cited by this paper
• Least Squares Generative Adversarial Networks

  2016cited by this paper
• Playing for Data: Ground Truth from Computer Games

  2016cited by this paper
• DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs

  2016influential reference
• Pyramid Scene Parsing Network

  2016cited by this paper
• FCNs in the Wild: Pixel-level Adversarial and Constraint-based Adaptation

  2016cited by this paper
• Temporal Ensembling for Semi-Supervised Learning

  2016cited by this paper
• The Cityscapes Dataset for Semantic Urban Scene Understanding

  2016cited by this paper
• Deep Residual Learning for Image Recognition

  2015cited by this paper
• Domain-Adversarial Training of Neural Networks

  2015cited by this paper
• Deep neural networks are easily fooled: High confidence predictions for unrecognizable images

  2014cited by this paper
• Explaining and Harnessing Adversarial Examples

  2014cited by this paper
• Fully convolutional networks for semantic segmentation

  2014influential reference
• Very Deep Convolutional Networks for Large-Scale Image Recognition

  2014influential reference
• Co-Training for Domain Adaptation

  2011cited by this paper
• Visualizing Data using t-SNE

  2008cited by this paper
• Subspace Pursuit for Compressive Sensing Signal Reconstruction

  2008cited by this paper
• Sparse Representation for Signal Classification

  2006cited by this paper
• Tri-training: exploiting unlabeled data using three classifiers

  2005cited by this paper
• Combining labeled and unlabeled data with co-training

  1998cited by this paper
• Summary of the

  1994cited by this paper
• Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper

  1977cited by this paper

• Modeling Image Segmentation Label Noise Through Concave Region Removal

  2026cites this paper
• Dual form Complementary Masking for Domain-Adaptive Image Segmentation

  2025cites this paper
• Hierarchical Prototype Learning With Uncertainty-Aware Adaptation for Cross-Domain Semantic Segmentation of Remote Sensing Images

  2025cites this paper
• Recent Advances in Deep Domain Adaptation Research for Semantic Segmentation in Urban Scenes

  2025cites this paper
• Preserving Label-Related Domain-Specific Information for Cross-Domain Semantic Segmentation

  2024cites this paper
• PDA: Progressive Domain Adaptation for Semantic Segmentation

  2024cites this paper
• JFDI: Joint Feature Differentiation and Interaction for domain adaptive object detection

  2024cites this paper
• Saliency-Free and Aesthetic-Aware Panoramic Video Navigation

  2024cites this paper
• Dual-level Interaction for Domain Adaptive Semantic Segmentation

  2023cites this paper
• Unified Domain Adaptive Semantic Segmentation

  2023cites this paper
• Robust Unsupervised Domain Adaptation by Retaining Confident Entropy via Edge Concatenation

  2023cites this paper
• Stable Prototype-Guided Single-Temporal Supervised Learning for Change Detection and Extraction of Building

  2023cites this paper
• Modified DSAN for unsupervised cross-domain fault diagnosis of bearing under speed fluctuation

  2022cites this paper