Towards Robust Pseudo-Label Learning in Semantic Segmentation: An Encoding Perspective

2025/12/07 by Li, Wangkai, Sun, Rui, Li, Zhaoyang +1
Computer Science · #Advanced Neural Network Applications #Computer Vision and Pattern Recognition (cs.CV) #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Machine Learning and Data Classification

paper · doi:10.48550/arxiv.2512.06870

openalex publication_date 2025/12/07 · openalex created_date 2025/12/10 · openalex updated_date 2026/07/28

Abstract

Pseudo-label learning is widely used in semantic segmentation, particularly in label-scarce scenarios such as unsupervised domain adaptation (UDA) and semisupervised learning (SSL). Despite its success, this paradigm can generate erroneous pseudo-labels, which are further amplified during training due to utilization of one-hot encoding. To address this issue, we propose ECOCSeg, a novel perspective for segmentation models that utilizes error-correcting output codes (ECOC) to create a fine-grained encoding for each class. ECOCSeg offers several advantages. First, an ECOC-based classifier is introduced, enabling model to disentangle classes into attributes and handle partial inaccurate bits, improving stability and generalization in pseudo-label learning. Second, a bit-level label denoising mechanism is developed to generate higher-quality pseudo-labels, providing adequate and robust supervision for unlabeled images. ECOCSeg can be easily integrated with existing methods and consistently demonstrates significant improvements on multiple UDA and SSL benchmarks across different segmentation architectures. Code is available at https://github.com/Woof6/ECOCSeg.

Citations

• Balanced Learning for Domain Adaptive Semantic Segmentation
• Target Semantics Clustering via Text Representations for Robust Universal Domain Adaptation
• Towards Unsupervised Domain Bridging via Image Degradation in Semantic Segmentation
• UniMatch V2: Pushing the Limit of Semi-Supervised Semantic Segmentation
• UniMatch V2: Pushing the Limit of Semi-Supervised Semantic Segmentation
• Localization and Expansion: A Decoupled Framework for Point Cloud Few-shot Semantic Segmentation
• Beyond Pixels: Semi-Supervised Semantic Segmentation with a Multi-scale Patch-based Multi-Label Classifier
• Image-to-Image Matching via Foundation Models: A New Perspective for Open-Vocabulary Semantic Segmentation
• AllSpark: Reborn Labeled Features from Unlabeled in Transformer for Semi-Supervised Semantic Segmentation
• CorrMatch: Label Propagation via Correlation Matching for Semi-Supervised Semantic Segmentation
• DINOv2: Learning Robust Visual Features without Supervision
• Semi-Supervised Semantic Segmentation via Gentle Teaching Assistant
• Augmentation Matters: A Simple-yet-Effective Approach to Semi-supervised Semantic Segmentation
• MIC: Masked Image Consistency for Context-Enhanced Domain Adaptation
• Fuzzy Positive Learning for Semi-supervised Semantic Segmentation
• Semi-supervised Semantic Segmentation with Prototype-based Consistency Regularization
• Learning from Future: A Novel Self-Training Framework for Semantic Segmentation
• Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic Segmentation
• Prototypical Contrast Adaptation for Domain Adaptive Semantic Segmentation
• FreeMatch: Self-adaptive Thresholding for Semi-supervised Learning
• Revisiting Consistency Regularization for Semi-supervised Change Detection in Remote Sensing Images
• Class-Balanced Pixel-Level Self-Labeling for Domain Adaptive Semantic Segmentation
• Smoothing Matters: Momentum Transformer for Domain Adaptive Semantic Segmentation
• Semi-Supervised Semantic Segmentation Using Unreliable Pseudo-Labels
• Semi-Supervised Medical Image Segmentation via Cross Teaching between CNN and Transformer
• Masked-attention Mask Transformer for Universal Image Segmentation
• Multiple Fusion Adaptation: A Strong Framework for Unsupervised Semantic Segmentation Adaptation
• DAFormer: Improving Network Architectures and Training Strategies for Domain-Adaptive Semantic Segmentation
• Perturbed and Strict Mean Teachers for Semi-supervised Semantic Segmentation
• Semi-Supervised Semantic Segmentation via Adaptive Equalization Learning
• Pixel Contrastive-Consistent Semi-Supervised Semantic Segmentation
• Multi-Anchor Active Domain Adaptation for Semantic Segmentation
• Learning Meta-class Memory for Few-Shot Semantic Segmentation
• Per-Pixel Classification is Not All You Need for Semantic Segmentation
• ST++: Make Self-training Work Better for Semi-supervised Semantic Segmentation
• Semi-Supervised Semantic Segmentation with Cross Pseudo Supervision
• Semi-supervised Semantic Segmentation with Directional Context-aware Consistency
• SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers
• Self-supervised Augmentation Consistency for Adapting Semantic Segmentation
• Prototypical Pseudo Label Denoising and Target Structure Learning for Domain Adaptive Semantic Segmentation
• An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
• PseudoSeg: Designing Pseudo Labels for Semantic Segmentation
• DACS: Domain Adaptation via Cross-domain Mixed Sampling
• Weakly-Supervised Semantic Segmentation via Sub-category Exploration
• Rectifying Pseudo Label Learning via Uncertainty Estimation for Domain Adaptive Semantic Segmentation
• FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
• PANet: Few-Shot Image Semantic Segmentation with Prototype Alignment
• Semi-Supervised Semantic Segmentation with High- and Low-level Consistency
• Uncertainty-aware Self-ensembling Model for Semi-supervised 3D Left Atrium Segmentation
• Fully Convolutional Networks for Multisource Building Extraction From an Open Aerial and Satellite Imagery Data Set
• ADVENT: Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentation
• Evidential Deep Learning to Quantify Classification Uncertainty
• Deep Learning Techniques for Automatic MRI Cardiac Multi-Structures Segmentation and Diagnosis: Is the Problem Solved?
• Learning Pixel-level Semantic Affinity with Image-level Supervision for Weakly Supervised Semantic Segmentation
• Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation
• Rethinking Atrous Convolution for Semantic Image Segmentation
• On Calibration of Modern Neural Networks
• Attention Is All You Need
• Pyramid Scene Parsing Network
• Xception: Deep Learning with Depthwise Separable Convolutions
• Playing for Data: Ground Truth from Computer Games
• Deep Reconstruction-Classification Networks for Unsupervised Domain Adaptation
• DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs
• DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs
• The Cityscapes Dataset for Semantic Urban Scene Understanding
• Deep Residual Learning for Image Recognition
• U-Net: Convolutional Networks for Biomedical Image Segmentation
• Fully Convolutional Networks for Semantic Segmentation
• Microsoft COCO: Common Objects in Context
• Distributed Representations of Words and Phrases and their Compositionality
• ImageNet classification with deep convolutional neural networks

Related

• PseudoSeg: Designing Pseudo Labels for Semantic Segmentation
• Pseudo-Label Noise Suppression Techniques for Semi-Supervised Semantic Segmentation
• Using Unreliable Pseudo-Labels for Label-Efficient Semantic Segmentation
• Zero-Shot Pseudo Labels Generation Using SAM and CLIP for Semi-Supervised Semantic Segmentation
• Pseudo Labeling Methods for Semi-Supervised Semantic Segmentation: A Review and Future Perspectives