Adversarial attacks on deep learning models for fatty liver disease classification by modification of ultrasound image reconstruction method

TL;DR

This paper introduces a novel adversarial attack on ultrasound-based deep learning models for fatty liver disease diagnosis, by perturbing image reconstruction parameters, highlighting vulnerabilities in medical imaging AI systems.

Contribution

The study presents a new adversarial attack method targeting ultrasound image reconstruction parameters to deceive deep learning models in medical diagnosis.

Findings

Attack success rate of 48% on fatty liver disease model

Perturbing reconstruction parameters can significantly alter model outputs

Ultrasound image reconstruction methods are vulnerable to adversarial manipulation

Abstract

Convolutional neural networks (CNNs) have achieved remarkable success in medical image analysis tasks. In ultrasound (US) imaging, CNNs have been applied to object classification, image reconstruction and tissue characterization. However, CNNs can be vulnerable to adversarial attacks, even small perturbations applied to input data may significantly affect model performance and result in wrong output. In this work, we devise a novel adversarial attack, specific to ultrasound (US) imaging. US images are reconstructed based on radio-frequency signals. Since the appearance of US images depends on the applied image reconstruction method, we explore the possibility of fooling deep learning model by perturbing US B-mode image reconstruction method. We apply zeroth order optimization to find small perturbations of image reconstruction parameters, related to attenuation compensation and amplitude compression, which can result in wrong output. We illustrate our approach using a deep learning model developed for fatty liver disease diagnosis, where the proposed adversarial attack achieved success rate of 48%.