Abstract:Although domain adaptation approaches have been proposed to tackle cross-regional, multitemporal, and multisensor remote sensing applications since they do not require any human interpretation in the target domain, most current works assume identical label space across the source and the target domains. However, in real-world applications, we often transfer knowledge from a large-scale dataset with rich annotations to a small-scale target dataset with scarcity of labels. In most cases, the label space of the source domain is usually large enough to subsume that of the target domain, which is termed partial domain adaptation. In this article, we propose a new partial domain adaptation algorithm for remote sensing scene classification and our proposed method contains three major parts. First, we employ a progressive auxiliary domain module to alleviate the negative transfer effect caused by outlier classes. Second, we adopt an improved domain adversarial neural network (DANN) with multiweights to better encourage domain confusion. Last but not least, we design an attentive complement entropy regularization to improve the prediction confidence for samples and avoid untransferable samples (such as the samples belonging to outlier classes in the source domain) being mistakenly classified. We collect three common remote sensing datasets to evaluate our proposed method. Our method achieves an average accuracy of 79.36%, which considerably outperforms other state-of-the-art partial domain adaptation methods with an average accuracy improvement of 1.90%–12.45% and attaining a 13.67% gain compared to the straightforward deep learning model (ResNet-50). The experiment results indicate that our approach shows promising prospects for solving more general and practical domain adaptation problems where the label space of the source domain subsumes that of the target domain.

Continual domain adaptation on aerial images under gradually degrading weather

Continual Domain Adaptation on Aerial Images under Gradually Degrading Weather

Multiscale Attention-Based Subdomain Dynamic Adaptation for Cross-Domain Scene Classification

Domain Adaptation on Point Clouds Via Geometry-Aware Implicits

Gradient Regularized Contrastive Learning for Continual Domain Adaptation

Continual Test-time Domain Adaptation via Dynamic Sample Selection

Domain Adaptation for Satellite-Borne Multispectral Cloud Detection

Domain Adaptation for Remote Sensing Image Semantic Segmentation: An Integrated Approach of Contrastive Learning and Adversarial Learning

Two-Stage Domain Adaptation Based on Image and Feature Levels for Cloud Detection in Cross-Spatiotemporal Domain

MsDA: Multi-scale Domain Adaptation Dehazing Network

Delving into the Continuous Domain Adaptation

ConDA: Continual Adaptation in Remote Sensing Via Visual Style Playback

DA-RAW: Domain Adaptive Object Detection for Real-World Adverse Weather Conditions

Source-Free Domain Adaptation for Real-World Image Dehazing

Partial Domain Adaptation for Scene Classification From Remote Sensing Imagery

Active Gradual Domain Adaptation: Dataset and Approach

A Step-Wise Domain Adaptation Detection Transformer for Object Detection under Poor Visibility Conditions

Complementary Domain Adaptation and Generalization for Unsupervised Continual Domain Shift Learning

Gradual Domain Adaptation: Theory and Algorithms

Progressive Domain Adaptation for Change Detection Using Season-Varying Remote Sensing Images

Continual Domain Adversarial Adaptation via Double-Head Discriminators