Source-Free Domain Adaptive Semantic Segmentation of Remote Sensing Images with Diffusion-Guided Label Enrichment

2025-09-22

AI Summary

This work addresses source-free domain adaptation (SFDA) for semantic segmentation of remote sensing images, where source-domain data are inaccessible during adaptation. We propose a diffusion-model-based framework for pseudo-label optimization and expansion. Our method introduces a controllable propagation mechanism initialized from high-quality seed pseudo-labels—avoiding noise accumulation inherent in full-image pseudo-label refinement. It integrates confidence-based filtering, super-resolution enhancement, and diffusion-prior modeling to construct a robust pseudo-label generation and fusion pipeline. The approach significantly improves segmentation accuracy on multiple remote sensing benchmarks, demonstrating superior pseudo-label quality and enhanced stability in self-training. To our knowledge, this is the first work to incorporate diffusion models into the SFDA paradigm for remote sensing segmentation.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Semi-Supervised Learning
• Search and Optimization: Learning to Search

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Research on unsupervised domain adaptation (UDA) for semantic segmentation of remote sensing images has been extensively conducted. However, research on how to achieve domain adaptation in practical scenarios where source domain data is inaccessible namely, source-free domain adaptation (SFDA) remains limited. Self-training has been widely used in SFDA, which requires obtaining as many high-quality pseudo-labels as possible to train models on target domain data. Most existing methods optimize the entire pseudo-label set to obtain more supervisory information. However, as pseudo-label sets often contain substantial noise, simultaneously optimizing all labels is challenging. This limitation undermines the effectiveness of optimization approaches and thus restricts the performance of self-training. To address this, we propose a novel pseudo-label optimization framework called Diffusion-Guided Label Enrichment (DGLE), which starts from a few easily obtained high-quality pseudo-labels and propagates them to a complete set of pseudo-labels while ensuring the quality of newly generated labels. Firstly, a pseudo-label fusion method based on confidence filtering and super-resolution enhancement is proposed, which utilizes cross-validation of details and contextual information to obtain a small number of high-quality pseudo-labels as initial seeds. Then, we leverage the diffusion model to propagate incomplete seed pseudo-labels with irregular distributions due to its strong denoising capability for randomly distributed noise and powerful modeling capacity for complex distributions, thereby generating complete and high-quality pseudo-labels. This method effectively avoids the difficulty of directly optimizing the complete set of pseudo-labels, significantly improves the quality of pseudo-labels, and thus enhances the model's performance in the target domain.

Problem

• Achieving domain adaptation without accessing source domain data for remote sensing image segmentation
• Optimizing noisy pseudo-labels in source-free domain adaptation for semantic segmentation
• Generating complete high-quality pseudo-labels from limited reliable seed labels

Innovation

• Uses diffusion models to propagate high-quality pseudo-labels
• Starts with a few seed labels from confidence filtering
• Generates complete, high-quality labels to avoid noise