FSDA-DETR: Few-Shot Domain-Adaptive Object Detection Transformer in Remote Sensing Imagery

Binbin Yang, Jianhong Han, Xinghai Hou, Dehao Zhou, Wenkai Liu, Fukun Bi

Abstract: Few-shot domain-adaptive object detection (FSDAOD) aims to transfer knowledge from a source domain to a target domain with limited labeled data, which faces severe challenges in the field of remote sensing. To address this issue, numerous convolutional neural network (CNN)-based domain adaptation methods employ style transfer and feature alignment to mitigate domain shifts, but limited target domain samples are prone to yielding equivocal optimization and are maladaptive. Furthermore, the sparsity of targets and the complexity of backgrounds in remote sensing imagery contribute to confusing feature alignment. Moreover, detection transformer (DETR)-based detectors have achieved remarkable progress in unsupervised domain adaptation (UDA) but remain unexplored in FSDAOD. To address these challenges, we introduce FSDA-DETR, the first DETR-based strong baseline designed for the FSDAOD of remote sensing imagery. Specifically, we propose a cross-domain style rectification (CSR) module that rectifies the styles of the target domain to align with the source domain by storing and dynamically updating the weighted-fusion source-domain style bases. To further strengthen the detector’s cross-domain detection performance, we propose a category-aware feature alignment (CFA) module that performs fine-grained masking on object regions of rectified backbone features corresponding to different categories and utilizes adversarial training for domain-invariant feature extraction. Extensive experiments on three cross-domain benchmarks, comprising six diverse datasets, demonstrate that FSDA-DETR outperforms state-of-the-art methods. For instance, in the optical-to-synthetic aperture radar (SAR) benchmark, FSDA-DETR achieves 74.5% mean average precision (mAP) with only 1% of target-domain training data. The code and datasets are available at https://github.com/wsybb252237/FSDA-DETR

External IDs:dblp:journals/tgrs/YangHHZLB25