Local Pixel-Contrast and Global Gaussian Multiprototype Bidirectional Alignment for Unsupervised Domain Adaptation of Semantic Segmentation in Remote Sensing Imagery

TY - JOUR

T1 - Local Pixel-Contrast and Global Gaussian Multiprototype Bidirectional Alignment for Unsupervised Domain Adaptation of Semantic Segmentation in Remote Sensing Imagery

AU - Hu, Yongkang

AU - Wang, Yupei

AU - Chen, Liang

N1 - Publisher Copyright: © 2008-2012 IEEE.

PY - 2026

Y1 - 2026

N2 - In unsupervised domain adaptation for semantic segmentation of remote sensing imagery, identical land-cover classes across different domains often exhibit substantial variations in appearance, scale, and class distribution, which seriously hinder cross-domain generalization. Moreover, even within a single domain, land-cover classes present highly complex and diverse intraclass distributions that cannot be effectively captured by a single class representation, further increasing the challenge of generalization. To this end, we propose a collaborative framework integrating local pixel-level contrast and global Gaussian multiprototype bidirectional alignment. At the local level, we introduce probability-masked contrastive learning, which adaptively increases the sampling probability of minority classes to mitigate the class imbalance issue. Meanwhile, pixel contrastive learning is incorporated to enhance the robustness to cross-domain variations in appearance and texture. At the global level, we employ a Gaussian mixture model to represent each source-domain class with multiple Gaussian prototypes rather than a single one, thereby yielding richer and more fine-grained class representations. Building on this, a bidirectional alignment strategy is proposed. Concretely, the forward alignment serves multiprototypes as semantic anchors that progressively guide target-domain features to align with the source-domain class distributions, reducing the intraclass variance. Meanwhile, the reverse alignment dynamically refines the prototypes to increase anchor accuracy, further enhancing the stability and discriminability of cross-domain alignment. Experimental results on the widely used ISPRS and LoveDA datasets demonstrate the superiority of our proposed method over state-of-the-art approaches.

AB - In unsupervised domain adaptation for semantic segmentation of remote sensing imagery, identical land-cover classes across different domains often exhibit substantial variations in appearance, scale, and class distribution, which seriously hinder cross-domain generalization. Moreover, even within a single domain, land-cover classes present highly complex and diverse intraclass distributions that cannot be effectively captured by a single class representation, further increasing the challenge of generalization. To this end, we propose a collaborative framework integrating local pixel-level contrast and global Gaussian multiprototype bidirectional alignment. At the local level, we introduce probability-masked contrastive learning, which adaptively increases the sampling probability of minority classes to mitigate the class imbalance issue. Meanwhile, pixel contrastive learning is incorporated to enhance the robustness to cross-domain variations in appearance and texture. At the global level, we employ a Gaussian mixture model to represent each source-domain class with multiple Gaussian prototypes rather than a single one, thereby yielding richer and more fine-grained class representations. Building on this, a bidirectional alignment strategy is proposed. Concretely, the forward alignment serves multiprototypes as semantic anchors that progressively guide target-domain features to align with the source-domain class distributions, reducing the intraclass variance. Meanwhile, the reverse alignment dynamically refines the prototypes to increase anchor accuracy, further enhancing the stability and discriminability of cross-domain alignment. Experimental results on the widely used ISPRS and LoveDA datasets demonstrate the superiority of our proposed method over state-of-the-art approaches.

KW - Contrastive learning

KW - prototype alignment

KW - unsupervised domain adaptation (UDA) of remote sensing images

UR - https://www.scopus.com/pages/publications/105025725954

U2 - 10.1109/JSTARS.2025.3647100

DO - 10.1109/JSTARS.2025.3647100

M3 - Article

AN - SCOPUS:105025725954

SN - 1939-1404

VL - 19

SP - 3573

EP - 3588

JO - IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing

JF - IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing

ER -