Source-Free Online Domain Adaptive Semantic Segmentation of Satellite Images under Image Degradation (2401.02113v1)
Abstract: Online adaptation to distribution shifts in satellite image segmentation stands as a crucial yet underexplored problem. In this paper, we address source-free and online domain adaptation, i.e., test-time adaptation (TTA), for satellite images, with the focus on mitigating distribution shifts caused by various forms of image degradation. Towards achieving this goal, we propose a novel TTA approach involving two effective strategies. First, we progressively estimate the global Batch Normalization (BN) statistics of the target distribution with incoming data stream. Leveraging these statistics during inference has the ability to effectively reduce domain gap. Furthermore, we enhance prediction quality by refining the predicted masks using global class centers. Both strategies employ dynamic momentum for fast and stable convergence. Notably, our method is backpropagation-free and hence fast and lightweight, making it highly suitable for on-the-fly adaptation to new domain. Through comprehensive experiments across various domain adaptation scenarios, we demonstrate the robust performance of our method.
- “S&gda: An unsupervised domain adaptive semantic segmentation framework considering both imaging scene and geometric domain shifts,” IEEE Transactions on Geoscience and Remote Sensing, 2023.
- “Standardgan: Multi-source domain adaptation for semantic segmentation of very high resolution satellite images by data standardization,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, June 2020.
- “Universal domain adaptation for remote sensing image scene classification,” IEEE Transactions on Geoscience and Remote Sensing, vol. 61, pp. 1–15, 2023.
- “Source-free domain adaptation for cross-scene hyperspectral image classification,” in IGARSS 2022-2022 IEEE International Geoscience and Remote Sensing Symposium. IEEE, 2022, pp. 3576–3579.
- “Tent: Fully test-time adaptation by entropy minimization,” in International Conference on Learning Representations, 2021.
- “Revisiting batch normalization for practical domain adaptation,” arXiv preprint arXiv:1603.04779, 2016.
- “Domain adaptation based on correlation subspace dynamic distribution alignment for remote sensing image scene classification,” IEEE Transactions on Geoscience and Remote Sensing, vol. 58, no. 11, pp. 7920–7930, 2020.
- “Unsupervised domain adaptation for remote sensing image segmentation based on adversarial learning and self-training,” IEEE Geoscience and Remote Sensing Letters, vol. 20, pp. 1–5, 2023.
- “The norm must go on: dynamic unsupervised domain adaptation by normalization,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 14765–14775.
- “Dynamically instance-guided adaptation: A backward-free approach for test-time domain adaptive semantic segmentation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 24090–24099.
- “Prototypical networks for few-shot learning,” Advances in neural information processing systems, vol. 30, 2017.
- “Transferrable prototypical networks for unsupervised domain adaptation,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 2239–2247.
- “Deepglobe 2018: A challenge to parse the earth through satellite images,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, 2018, pp. 172–181.
- “Robustbench: a standardized adversarial robustness benchmark,” in Thirty-fifth Conference on Neural Information Processing Systems Datasets and Benchmarks Track (Round 2), 2021.
- “Rethinking atrous convolution for semantic image segmentation,” arXiv preprint arXiv:1706.05587, 2017.
- Fahim Faisal Niloy (12 papers)
- Kishor Kumar Bhaumik (8 papers)
- Simon S. Woo (42 papers)