Semantic segmentation models based on convolutional neural networks (CNNs) have gained much attention in relation to remote sensing and have achieved remarkable performance for the extraction of buildings from high-resolution aerial images. However, the issue of limited generalization for unseen images remains. When there is a domain gap between the training and test datasets, CNN-based segmentation models trained by a training dataset fail to segment buildings for the test dataset. In this paper, we propose segmentation networks based on a domain adaptive transfer attack (DATA) scheme for building extraction from aerial images. The proposed system combines the domain transfer and adversarial attack concepts. Based on the DATA scheme, the distribution of the input images can be shifted to that of the target images while turning images into adversarial examples against a target network. Defending adversarial examples adapted to the target domain can overcome the performance degradation due to the domain gap and increase the robustness of the segmentation model. Cross-dataset experiments and the ablation study are conducted for the three different datasets: the Inria aerial image labeling dataset, the Massachusetts building dataset, and the WHU East Asia dataset. Compared to the performance of the segmentation network without the DATA scheme, the proposed method shows improvements in the overall IoU. Moreover, it is verified that the proposed method outperforms even when compared to feature adaptation (FA) and output space adaptation (OSA).
Table Of Contents
Ⅰ. Introduction 1 1.1. Semantic segmentation of aerial images 1 1.2. Challenging issue in CNN-based segmentation models 2 1.3. Overview of proposed scheme 2 Ⅱ. Basic architecture for semantic segmentation 4 2.1. Semantic segmentation systems 4 2.2. Inria aerial image labeling dataset 6 2.3. Training setup 6 2.4. Test results and comparison with other architectures 7 2.5. Test results with other datasets 8 2.6. Related work for domain adaptation 11 Ⅲ. Domain adaptive transfer attack (DATA) 11 3.1. Overview of the proposed model 12 3.2. Objective function for the generator 14 3.3. Objective function for discriminator 15 3.4. Training the adversarial attack model & discriminator 16 Ⅳ. DATA-based adversarial training and results 19 4.1. Adversary training setup 19 4.2. Comparison with Other Methods 21 4.3. Extended experiments in various environments 24 Ⅴ. Conclusion 26 References 28