Adversarial Test on Learnable Image Encryption
This work addresses privacy concerns in deep learning by evaluating the security of encryption methods against adversarial threats, though it appears incremental as it tests existing approaches without proposing new solutions.
The paper investigated the vulnerability of learnable image encryption to adversarial attacks across five scenarios, finding that it offers some adversarial robustness but with varying network behaviors depending on key settings.
Data for deep learning should be protected for privacy preserving. Researchers have come up with the notion of learnable image encryption to satisfy the requirement. However, existing privacy preserving approaches have never considered the threat of adversarial attacks. In this paper, we ran an adversarial test on learnable image encryption in five different scenarios. The results show different behaviors of the network in the variable key scenarios and suggest learnable image encryption provides certain level of adversarial robustness.