Bridging the Gap between Labeled and Unlabeled Data via Unified Flow with Feature Memory Bank

2026-08-17

AI Summary

This study addresses the challenges of label bias-induced pseudo-label degradation and feature misalignment in semi-supervised remote sensing segmentation. We propose UFFM, a Unified Flow Framework that integrates visual foundation models with domain-specific teachers to generate unbiased pseudo-labels within a unified training paradigm. Furthermore, a dynamic feature memory bank is designed to align class-level features, enabling joint optimization of labeled and pseudo-labeled data. Extensive experiments demonstrate that UFFM outperforms state-of-the-art methods across multiple remote sensing datasets. By effectively bridging the gap between optimization objectives and feature representations, this work establishes a novel paradigm for semi-supervised learning in remote sensing applications.

Abstract

Although semi-supervised semantic segmentation ($\text{S}^4$) utilizes abundant unlabeled data to reduce manual labeling burdens, independent training of labeled and unlabeled data causes the former to dominate, which severely degrades pseudo-label quality. To address this challenges, we propose a novel remote sensing (RS) $\text{S}^4$ method via unified flow with feature memory bank (UFFM). Specifically, UFFM comprises two key innovations: unified flow (UF) and feature memory bank (FMB). The UF is a new training flow that generates less biased pseudo-labels by combining an external visual foundation model (VFM) with an RS domain teacher, and jointly optimizes labeled and pseudo-labeled data under a unified training objective. The FMB is a novel memory module for $\text{S}^4$ that dynamically updates class-specific features during training and reduces the feature discrepancy between labeled and unlabeled data through class-feature alignment. To verify the effectiveness of our model, we conduct extensive experiments on RS datasets. The experimental results show the superiority of our method over SOTA $\text{S}^4$ methods. Moreover, the results demonstrate the effectiveness of our contributions in bridging the optimization and feature representation gap between labeled and unlabeled data. Our code is released at \href{https://github.com/wangshanwen001/RS-UFFM}{https://github.com/wangshanwen001/RS-UFFM}.

Problem

• Semi-supervised Semantic Segmentation
• Remote Sensing
• Pseudo-label Quality
• Feature Discrepancy

Innovation

• Unified Flow
• Feature Memory Bank
• Semi-supervised Semantic Segmentation
• Visual Foundation Model
• Remote Sensing