EGPO: Enhanced Guidance and Pseudo-Label Optimization for Semi-Supervised Semantic Segmentation of Remote Sensing Images

Owing to complex remote sensing image features boosting manually labeled costs, semi-supervised semantic segmentation learns from limited labeled and abundant unlabeled data to alleviate this dilemma. However, there are still many challenges in the practical application of this technology, such as incorrect unsupervised information misleading the model and causing errors in pseudo-labels, thereby causing error accumulation. This article proposes a semi-supervised semantic segmentation method for remote sensing images. The enhanced guided learning module we designed uses a label-guided model to predict unlabeled data in the correct direction, enriching ground features and unsupervised information and alleviating the problems of misprediction and consistency regularization failure caused by the lack of labels. At the same time, facing the noise generated during the data augmentation process, our designed unsupervised loss dynamic screening module aims to locate and suppress the noise adaptively. In addition, in the face of inevitable erroneous predictions in pseudo-labels, we design a pixel category selection (PCS) module that produces high-quality and high-confidence pseudo-labels through multistep filtering and dual model fusion. Ultimately, by conducting experiments on the DFC22, iSAID, MER, GID-15, and Vaihingen datasets, we successfully verified the effectiveness of our proposed method. The source code has been made public: https://github.com/Xidian-AIGroup190726/EGPO

• Publication year

  2025

• Venue

  IEEE Transactions on Geoscience and Remote Sensing

• Publication date

  Unknown publication date

• Fields of study

  Computer Science, Environmental Science

• Identifiers
  DOI 10.1109/TGRS.2025.3630639
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Negative Class Guided Spatial Consistency Network for Sparsely Supervised Semantic Segmentation of Remote Sensing Images

  2025cited by this paper
• Incorporating Multiscale Context and Task-Consistent Focal Loss into Oriented Object Detection

  2025cited by this paper
• Scale matters: How spatial resolution impacts remote sensing based urban green space mapping?

  2024influential reference
• A Semantically Nonredundant Continuous-Scale Feature Network for Panchromatic and Multispectral Classification

  2024cited by this paper
• Few-Shot MS and PAN Joint Classification With Improved Cross-Source Contrastive Learning

  2024cited by this paper
• Complete and Invariant Instance Classifier Refinement for Weakly Supervised Object Detection in Remote Sensing Images

  2024cited by this paper
• Dynamic and Adaptive Self-Training for Semi-Supervised Remote Sensing Image Semantic Segmentation

  2024cited by this paper
• Modeling the Label Distributions for Weakly-Supervised Semantic Segmentation

  2024cited by this paper
• Self Pseudo Entropy Knowledge Distillation for Semi-Supervised Semantic Segmentation

  2024cited by this paper
• Enhanced Pseudo-Label Generation With Self-Supervised Training for Weakly- Supervised Semantic Segmentation

  2024cited by this paper
• Intra- and Intersource Interactive Representation Learning Network for Remote Sensing Images Classification

  2024cited by this paper
• Dual-Path Feature Aware Network for Remote Sensing Image Semantic Segmentation

  2024cited by this paper
• Semi-Supervised Few-Shot Object Detection via Adaptive Pseudo Labeling

  2024cited by this paper
• UniMatch V2: Pushing the Limit of Semi-Supervised Semantic Segmentation

  2024cited by this paper
• Semi-supervised Deep Learning via Transformation Consistency Regularization for Remote Sensing Image Semantic Segmentation

  2023cited by this paper
• Querying Labeled for Unlabeled: Cross-Image Semantic Consistency Guided Semi-Supervised Semantic Segmentation

  2023cited by this paper
• Spectral Super-Resolution via Model-Guided Cross-Fusion Network

  2023cited by this paper
• Building a Bridge of Bounding Box Regression Between Oriented and Horizontal Object Detection in Remote Sensing Images

  2023cited by this paper
• Weak-to-Strong Consistency Learning for Semisupervised Image Segmentation

  2023influential reference
• FarSeg++: Foreground-Aware Relation Network for Geospatial Object Segmentation in High Spatial Resolution Remote Sensing Imagery

  2023cited by this paper
• Cross-City Matters: A Multimodal Remote Sensing Benchmark Dataset for Cross-City Semantic Segmentation using High-Resolution Domain Adaptation Networks

  2023cited by this paper
• CMTFNet: CNN and Multiscale Transformer Fusion Network for Remote-Sensing Image Semantic Segmentation

  2023cited by this paper
• Semi-supervised Semantic Segmentation with Prototype-based Consistency Regularization

  2022cited by this paper
• A Stepwise Domain Adaptive Segmentation Network with Covariate shift Alleviation for Remote Sensing Imagery

  2022cited by this paper
• Semi-Supervised Semantic Segmentation Using Unreliable Pseudo-Labels

  2022cited by this paper
• ASS-GAN: Asymmetric semi-supervised GAN for breast ultrasound image segmentation

  2022cited by this paper
• The 2022 IEEE GRSS Data Fusion Contest: Semisupervised Learning [Technical Committees]

  2022cited by this paper
• Tensor Decompositions for Hyperspectral Data Processing in Remote Sensing: A comprehensive review

  2022cited by this paper
• Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic Segmentation

  2022cited by this paper
• Semi-Supervised Semantic Segmentation of Vessel Images using Leaking Perturbations

  2021cited by this paper
• ST++: Make Self-trainingWork Better for Semi-supervised Semantic Segmentation

  2021cited by this paper
• Semantic Segmentation of Remote Sensing Images With Sparse Annotations

  2021cited by this paper
• ClassMix: Segmentation-Based Data Augmentation for Semi-Supervised Learning

  2020influential reference
• A Three-Stage Self-Training Framework for Semi-Supervised Semantic Segmentation

  2020cited by this paper
• Semi-supervised Medical Image Segmentation through Dual-task Consistency

  2020cited by this paper
• iSAID: A Large-scale Dataset for Instance Segmentation in Aerial Images

  2019cited by this paper
• ResUNet-a: a deep learning framework for semantic segmentation of remotely sensed data

  2019cited by this paper
• Semi-Supervised Semantic Segmentation With High- and Low-Level Consistency

  2019cited by this paper
• CutMix: Regularization Strategy to Train Strong Classifiers With Localizable Features

  2019cited by this paper
• Adversarial Learning for Semi-supervised Semantic Segmentation

  2018cited by this paper
• Land-cover classification with high-resolution remote sensing images using transferable deep models

  2018cited by this paper
• Semi Supervised Semantic Segmentation Using Generative Adversarial Network

  2017cited by this paper
• Regularization With Stochastic Transformations and Perturbations for Deep Semi-Supervised Learning

  2016cited by this paper
• Pseudo-Label : The Simple and Efficient Semi-Supervised Learning Method for Deep Neural Networks

  2013cited by this paper

• Research on inversion and prediction of root region soil water content in kiwifruit based on hyperparameter tuning by transformer-DsaGRU

  2026cites this paper
• A Semi-Supervised Transformer with a Curriculum Training Pipeline for Remote Sensing Image Semantic Segmentation

  2026cites this paper
• Contrasting spatial and temporal patterns in chlorophyll-a and turbidity in Lake Okeechobee: A satellite perspective

  2026cites this paper