Adversarial Examples on Segmentation Models Can be Easy to Transfer
This work addresses the adversarial robustness problem for semantic segmentation models, which is incremental as it builds on prior studies of classification models.
The paper investigates the transferability of adversarial examples on semantic segmentation models, finding that unlike classification models, segmentation models do not always suffer from overfitting, but transferability is still limited due to multi-scale object recognition. It proposes a dynamic scaling method that achieves high transferability, demonstrating that adversarial examples on segmentation models can be easily transferred to other models.
Deep neural network-based image classification can be misled by adversarial examples with small and quasi-imperceptible perturbations. Furthermore, the adversarial examples created on one classification model can also fool another different model. The transferability of the adversarial examples has recently attracted a growing interest since it makes black-box attacks on classification models feasible. As an extension of classification, semantic segmentation has also received much attention towards its adversarial robustness. However, the transferability of adversarial examples on segmentation models has not been systematically studied. In this work, we intensively study this topic. First, we explore the overfitting phenomenon of adversarial examples on classification and segmentation models. In contrast to the observation made on classification models that the transferability is limited by overfitting to the source model, we find that the adversarial examples on segmentations do not always overfit the source models. Even when no overfitting is presented, the transferability of adversarial examples is limited. We attribute the limitation to the architectural traits of segmentation models, i.e., multi-scale object recognition. Then, we propose a simple and effective method, dubbed dynamic scaling, to overcome the limitation. The high transferability achieved by our method shows that, in contrast to the observations in previous work, adversarial examples on a segmentation model can be easy to transfer to other segmentation models. Our analysis and proposals are supported by extensive experiments.