Adversarial Robustness of Deep Learning Models for Inland Water Body Segmentation from SAR Images
This work addresses the impact of label noise from manual annotations on segmentation accuracy for applications like flood mapping, but it is incremental as it focuses on simulating errors rather than proposing a new robust method.
The paper tackled the problem of adversarial robustness in deep learning models for inland water body segmentation from SAR images, simulating manual annotation errors as adversarial attacks on a U-Net model, and found that the model can tolerate a certain level of corruption before performance drops significantly.
Inland water body segmentation from Synthetic Aperture Radar (SAR) images is an important task needed for several applications, such as flood mapping. While SAR sensors capture data in all-weather conditions as high-resolution images, differentiating water and water-like surfaces from SAR images is not straightforward. Inland water bodies, such as large river basins, have complex geometry, which adds to the challenge of segmentation. U-Net is a widely used deep learning model for land-water segmentation of SAR images. In practice, manual annotation is often used to generate the corresponding water masks as ground truth. Manual annotation of the images is prone to label noise owing to data poisoning attacks, especially due to complex geometry. In this work, we simulate manual errors in the form of adversarial attacks on the U-Net model and study the robustness of the model to human errors in annotation. Our results indicate that U-Net can tolerate a certain level of corruption before its performance drops significantly. This finding highlights the crucial role that the quality of manual annotations plays in determining the effectiveness of the segmentation model. The code and the new dataset, along with adversarial examples for robust training, are publicly available. (GitHub link - https://github.com/GVCL/IWSeg-SAR-Poison.git)