Biblio

In recent years, the security of AI systems has drawn increasing research attention, especially in the medical imaging realm. To develop a secure medical image analysis (MIA) system, it is a must to study possible backdoor attacks (BAs), which can embed hidden malicious behaviors into the system. However, designing a unified BA method that can be applied to various MIA systems is challenging due to the diversity of imaging modalities (e.g., X-Ray, CT, and MRI) and analysis tasks (e.g., classification, detection, and segmentation). Most existing BA methods are designed to attack natural image classification models, which apply spatial triggers to training images and inevitably corrupt the semantics of poisoned pixels, leading to the failures of attacking dense prediction models. To address this issue, we propose a novel Frequency-Injection based Backdoor Attack method (FIBA) that is capable of delivering attacks in various MIA tasks. Specifically, FIBA leverages a trigger function in the frequency domain that can inject the low-frequency information of a trigger image into the poisoned image by linearly combining the spectral amplitude of both images. Since it preserves the semantics of the poisoned image pixels, FIBA can perform attacks on both classification and dense prediction models. Experiments on three benchmarks in MIA (i.e., ISIC-2019 [4] for skin lesion classification, KiTS-19 [17] for kidney tumor segmentation, and EAD-2019 [1] for endoscopic artifact detection), validate the effectiveness of FIBA and its superiority over stateof-the-art methods in attacking MIA models and bypassing backdoor defense. Source code will be available at code.

This paper presents Themis, an end-to-end static analysis tool for finding resource-usage side-channel vulnerabilities in Java applications. We introduce the notion of epsilon-bounded non-interference, a variant and relaxation of Goguen and Meseguer's well-known non-interference principle. We then present Quantitative Cartesian Hoare Logic (QCHL), a program logic for verifying epsilon-bounded non-interference. Our tool, Themis, combines automated reasoning in CHL with lightweight static taint analysis to improve scalability. We evaluate Themis on well known Java applications and demonstrate that Themis can find unknown side-channel vulnerabilities in widely-used programs. We also show that Themis can verify the absence of vulnerabilities in repaired versions of vulnerable programs and that Themis compares favorably against Blazer, a state-of-the-art static analysis tool for finding timing side channels in Java applications.