mlsanityData to 14 Oct 2026

Semi-Supervised Semantic Segmentation via Gentle Teaching Assistant

arXiv 2301.07340 ↗First posted 18 Jan 2023cs.CVcited by 34

On the map: Knowledge distillation → Detection and dense prediction distillation

Abstract (arXiv)

Semi-Supervised Semantic Segmentation aims at training the segmentation model with limited labeled data and a large amount of unlabeled data. To effectively leverage the unlabeled data, pseudo labeling, along with the teacher-student framework, is widely adopted in semi-supervised semantic segmentation. Though proved to be effective, this paradigm suffers from incorrect pseudo labels which inevitably exist and are taken as auxiliary training data. To alleviate the negative impact of incorrect pseudo labels, we delve into the current Semi-Supervised Semantic Segmentation frameworks. We argue that the unlabeled data with pseudo labels can facilitate the learning of representative features in the feature extractor, but it is unreliable to supervise the mask predictor. Motivated by this consideration, we propose a novel framework, Gentle Teaching Assistant (GTA-Seg) to disentangle the effects of pseudo labels on feature extractor and mask predictor of the student model. Specifically, in addition to the original teacher-student framework, our method introduces a teaching assistant network which directly learns from pseudo labels generated by the teacher network. The gentle teaching assistant (GTA) is coined gentle since it only transfers the beneficial feature representation knowledge in the feature extractor to the student model in an Exponential Moving Average (EMA) manner, protecting the student model from the negative influences caused by unreliable pseudo labels in the mask predictor. The student model is also supervised by reliable labeled data to train an accurate mask predictor, further facilitating feature representation. Extensive experiment results on benchmark datasets validate that our method shows competitive performance against previous methods. Code is available at https://github.com/Jin-Ying/GTA-Seg.

Key facts. Cited by 34 papers in the map (data through 14 Oct 2026). Placed in Detection and dense prediction distillation, one of 12 directions in Knowledge distillation. Builds on 41 of its 60 references that are in the map. First posted on arXiv 18 Jan 2023.

Builds on

41 of 60 references.

References · 41

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2017On Calibration of Modern Neural Networkscited by 4,960
  3. 2014Fully Convolutional Networks for Semantic Segmentationcited by 4,859
  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. 2016Pyramid Scene Parsing Networkcited by 2,479
  8. 2020FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidencecited by 2,085
  9. 2015SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentationcited by 1,487
  10. 2019MixMatch: A Holistic Approach to Semi-Supervised Learningcited by 1,391
  11. 2019Deep High-Resolution Representation Learning for Human Pose Estimationcited by 1,339
  12. 2015Multi-Scale Context Aggregation by Dilated Convolutionscited by 1,133
29 more
  1. 2019Unsupervised Data Augmentation for Consistency Trainingcited by 1,060
  2. 2016Temporal Ensembling for Semi-Supervised Learningcited by 1,057
  3. 2018Dual Attention Network for Scene Segmentationcited by 1,003
  4. 2017Virtual Adversarial Training: A Regularization Method for Supervised and Semi-Supervised Learningcited by 920
  5. 2018CCNet: Criss-Cross Attention for Semantic Segmentationcited by 718
  6. 2021FlexMatch: Boosting Semi-Supervised Learning with Curriculum Pseudo Labelingcited by 540
  7. 2021Semi-Supervised Semantic Segmentation with Cross Pseudo Supervisioncited by 463
  8. 2016Regularization With Stochastic Transformations and Perturbations for Deep Semi-Supervised Learningcited by 441
  9. 2020Semi-Supervised Semantic Segmentation with Cross-Consistency Trainingcited by 376
  10. 2015Learning Deconvolution Network for Semantic Segmentationcited by 365
  11. 2021End-to-End Semi-Supervised Object Detection with Soft Teachercited by 316
  12. 2018Context Encoding for Semantic Segmentationcited by 290
  13. 2022Semi-Supervised Semantic Segmentation Using Unreliable Pseudo-Labelscited by 218
  14. 2020ClassMix: Segmentation-Based Data Augmentation for Semi-Supervised Learningcited by 209
  15. 2021ST++: Make Self-training Work Better for Semi-supervised Semantic Segmentationcited by 184
  16. 2020Theoretical Analysis of Self-Training with Deep Networks on Unlabeled Datacited by 180
  17. 2020PseudoSeg: Designing Pseudo Labels for Semantic Segmentationcited by 180
  18. 2015ParseNet: Looking Wider to See Bettercited by 177
  19. 2018Adversarial Learning for Semi-Supervised Semantic Segmentationcited by 173
  20. 2019Semi-Supervised Semantic Segmentation with High- and Low-level Consistencycited by 147
  21. 2021Semi-Supervised Semantic Segmentation with Pixel-Level Contrastive Learning from a Class-wise Memory Bankcited by 125
  22. 2020Guided Collaborative Training for Pixel-wise Semi-Supervised Learningcited by 120
  23. 2021Semi-supervised Semantic Segmentation with Directional Context-aware Consistencycited by 112
  24. 2021Pixel Contrastive-Consistent Semi-Supervised Semantic Segmentationcited by 101
  25. 2021Semi-Supervised Semantic Segmentation via Adaptive Equalization Learningcited by 101
  26. 2021A Simple Baseline for Semi-supervised Semantic Segmentation with Strong Data Augmentationcited by 79
  27. 2020DMT: Dynamic Mutual Training for Semi-Supervised Learningcited by 77
  28. 2018Semi-Supervised Semantic Image Segmentation with Self-correcting Networkscited by 37
  29. 2017Unsupervised learning from video to detect foreground objects in single imagescited by 15

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 8; 2024: 10; 2025: 13; 2026: 3.

34 papers.

Sort:

By area:

  1. 2024UniMatch V2: Pushing the Limit of Semi-Supervised Semantic Segmentationcited by 54
  2. 2023CorrMatch: Label Propagation via Correlation Matching for Semi-Supervised Semantic Segmentationcited by 44
  3. 2024AllSpark: Reborn Labeled Features from Unlabeled in Transformer for Semi-Supervised Semantic Segmentationcited by 35
  4. 2023Logic-induced Diagnostic Reasoning for Semi-supervised Semantic Segmentationcited by 29
  5. 2023Switching Temporary Teachers for Semi-Supervised Semantic Segmentationcited by 23
  6. 2024Towards the Uncharted: Density-Descending Feature Perturbation for Semi-supervised Semantic Segmentationcited by 16
  7. 2024Semi-Supervised Semantic Segmentation Based on Pseudo-Labels: A Surveycited by 12
  8. 2023Semi-supervised Segmentation of Histopathology Images with Noise-Aware Topological Consistencycited by 11
  9. 2025Semi-supervised Semantic Segmentation for Remote Sensing Images via Multi-scale Uncertainty Consistency and Cross-Teacher-Student Attentioncited by 10
  10. 2023Bridging Synthetic and Real Worlds for Pre-training Scene Text Detectorscited by 9
  11. 2024PrevMatch: Revisiting and Maximizing Temporal Knowledge in Semi-Supervised Semantic Segmentationcited by 6
  12. 2025When Confidence Fails: Revisiting Pseudo-Label Selection in Semi-supervised Semantic Segmentationcited by 6
22 more
  1. 2023Improving Semi-Supervised Semantic Segmentation with Dual-Level Siamese Structure Networkcited by 4
  2. 2023Semi-supervised Semantic Segmentation Meets Masked Modeling:Fine-grained Locality Learning Matters in Consistency Regularizationcited by 4
  3. 2025ConformalSAM: Unlocking the Potential of Foundational Segmentation Models in Semi-Supervised Semantic Segmentation with Conformal Predictioncited by 4
  4. 2025Semi-MoE: Mixture-of-Experts meets Semi-Supervised Histopathology Segmentationcited by 4
  5. 2025DUDA: Distilled Unsupervised Domain Adaptation for Lightweight Semantic Segmentationcited by 3
  6. 2025Rethinking Semi-supervised Segmentation Beyond Accuracy: Reliability and Robustnesscited by 3
  7. 2023Boundary-Refined Prototype Generation: A General End-to-End Paradigm for Semi-Supervised Semantic Segmentationcited by 2
  8. 2024SepRep-Net: Multi-source Free Domain Adaptation via Model Separation And Reparameterizationcited by 2
  9. 2024A Point-Neighborhood Learning Framework for Nasal Endoscope Image Segmentationcited by 2
  10. 2025FARCLUSS: Fuzzy Adaptive Rebalancing and Contrastive Uncertainty Learning for Semi-Supervised Semantic Segmentationcited by 2
  11. 2025Towards Robust Pseudo-Label Learning in Semantic Segmentation: An Encoding Perspectivecited by 2
  12. 2024Boosting Semi-Supervised Temporal Action Localization by Learning from Non-Target Classescited by 1
  13. 2024Multi-Level Label Correction by Distilling Proximate Patterns for Semi-supervised Semantic Segmentationcited by 1
  14. 2025Knowledge Consultation for Semi-Supervised Semantic Segmentationcited by 1
  15. 2025Adaptive Spatial Augmentation for Semi-supervised Semantic Segmentationcited by 1
  16. 2025DuetMatch: Harmonizing Semi-Supervised Brain MRI Segmentation via Decoupled Branch Optimizationcited by 1
  17. 2024Augmenting Offline RL with Unlabeled Datacited by 0
  18. 2025Leveraging Out-of-Distribution Unlabeled Images: Semi-Supervised Semantic Segmentation with an Open-Vocabulary Modelcited by 0
  19. 2025Semi-Supervised Semantic Segmentation via Derivative Label Propagationcited by 0
  20. 2026CoVar: Confidence-Variance-Guided Pseudo-Label Selection for Semi-Supervised Learningcited by 0
  21. 2026WeatherSeg: Weather-Robust Image Segmentation using Teacher-Student Dual Learning and Classifier-Updating Attentioncited by 0
  22. 2026PixCon: Clean-Positive Contrastive Learning for Foundation-Model Semi-Supervised Segmentationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2022Polite Teacher: Semi-Supervised Instance Segmentation with Mutual Learning and Pseudo-Label Thresholding
  2. 2022Augmentation Matters: A Simple-yet-Effective Approach to Semi-supervised Semantic Segmentation
  3. 2024Better (pseudo-)labels for semi-supervised instance segmentation
  4. 2025S^4M: Boosting Semi-Supervised Instance Segmentation with SAM
  5. 2022Semantic Segmentation with Active Semi-Supervised Learning
  6. 2023SemiVL: Semi-Supervised Semantic Segmentation with Vision-Language Guidance
  7. 2022Semi-supervised Semantic Segmentation with Mutual Knowledge Distillation
  8. 2023Guided Distillation for Semi-Supervised Instance Segmentation