arXiv subjects

Junhao Dong

Publications and source records attributed to Junhao Dong.

65 records · Page 4Linked to original sources

arXiv2024

Survey on Adversarial Attack and Defense for Medical Image Analysis: Methods and Challenges

Deep learning techniques have achieved superior performance in computer-aided medical image analysis, yet they are still vulnerable to imperceptible adversarial attacks, resulting in potential misdiagnosis in clinical practice. Oppositely, recent years have also witnessed remarkable progress in defense against these tailored adversarial examples in deep medical diagnosis systems. In this exposition, we present a comprehensive survey on recent advances in adversarial attacks and defenses for medical image analysis with a systematic taxonomy in terms of the application scenario. We also provide a unified framework for different types of adversarial attack and defense methods in the context of medical image analysis. For a fair comparison, we establish a new benchmark for adversarially robust medical diagnosis models obtained by adversarial training under various scenarios. To the best of our knowledge, this is the first survey paper that provides a thorough evaluation of adversarially robust medical diagnosis models. By analyzing qualitative and quantitative results, we conclude this survey with a detailed discussion of current challenges for adversarial attack and defense in medical image analysis systems to shed light on future research directions. Code is available on \href{https://github.com/tomvii/Adv_MIA}{\color{red}{GitHub}}.

eess.IV↗

arXiv2024

Boundary-Refined Prototype Generation: A General End-to-End Paradigm for Semi-Supervised Semantic Segmentation

Semi-supervised semantic segmentation has attracted increasing attention in computer vision, aiming to leverage unlabeled data through latent supervision. To achieve this goal, prototype-based classification has been introduced and achieved lots of success. However, the current approaches isolate prototype generation from the main training framework, presenting a non-end-to-end workflow. Furthermore, most methods directly perform the K-Means clustering on features to generate prototypes, resulting in their proximity to category semantic centers, while overlooking the clear delineation of class boundaries. To address the above problems, we propose a novel end-to-end boundary-refined prototype generation (BRPG) method. Specifically, we perform online clustering on sampled features to incorporate the prototype generation into the whole training framework. In addition, to enhance the classification boundaries, we sample and cluster high- and low-confidence features separately based on confidence estimation, facilitating the generation of prototypes closer to the class boundaries. Moreover, an adaptive prototype optimization strategy is proposed to increase the number of prototypes for categories with scattered feature distributions, which further refines the class boundaries. Extensive experiments demonstrate the remarkable robustness and scalability of our method across diverse datasets, segmentation networks, and semi-supervised frameworks, outperforming the state-of-the-art approaches on three benchmark datasets: PASCAL VOC 2012, Cityscapes and MS COCO. The code is available at https://github.com/djh-dzxw/BRPG.

cs.CV↗

arXiv2024

Artwork Protection Against Neural Style Transfer Using Locally Adaptive Adversarial Color Attack

Neural style transfer (NST) generates new images by combining the style of one image with the content of another. However, unauthorized NST can exploit artwork, raising concerns about artists' rights and motivating the development of proactive protection methods. We propose Locally Adaptive Adversarial Color Attack (LAACA), empowering artists to protect their artwork from unauthorized style transfer by processing before public release. By delving into the intricacies of human visual perception and the role of different frequency components, our method strategically introduces frequency-adaptive perturbations in the image. These perturbations significantly degrade the generation quality of NST while maintaining an acceptable level of visual change in the original image, ensuring that potential infringers are discouraged from using the protected artworks, because of its bad NST generation quality. Additionally, existing metrics often overlook the importance of color fidelity in evaluating color-mattered tasks, such as the quality of NST-generated images, which is crucial in the context of artistic works. To comprehensively assess the color-mattered tasks, we propose the Adversarial Color Distance Metric (ACDM), designed to quantify the color difference of images pre- and post-manipulations. Experimental results confirm that attacking NST using LAACA results in visually inferior style transfer, and the ACDM can efficiently measure color-mattered tasks. By providing artists with a tool to safeguard their intellectual property, our work relieves the socio-technical challenges posed by the misuse of NST in the art community.

cs.CV↗

arXiv2024

XIMAGENET-12: An Explainable AI Benchmark Dataset for Model Robustness Evaluation

Despite the promising performance of existing visual models on public benchmarks, the critical assessment of their robustness for real-world applications remains an ongoing challenge. To bridge this gap, we propose an explainable visual dataset, XIMAGENET-12, to evaluate the robustness of visual models. XIMAGENET-12 consists of over 200K images with 15,410 manual semantic annotations. Specifically, we deliberately selected 12 categories from ImageNet, representing objects commonly encountered in practical life. To simulate real-world situations, we incorporated six diverse scenarios, such as overexposure, blurring, and color changes, etc. We further develop a quantitative criterion for robustness assessment, allowing for a nuanced understanding of how visual models perform under varying conditions, notably in relation to the background. We make the XIMAGENET-12 dataset and its corresponding code openly accessible at \url{https://sites.google.com/view/ximagenet-12/home}. We expect the introduction of the XIMAGENET-12 dataset will empower researchers to thoroughly evaluate the robustness of their visual models under challenging conditions.

cs.CV↗

arXiv2024

Exploring Adversarial Attacks against Latent Diffusion Model from the Perspective of Adversarial Transferability

Recently, many studies utilized adversarial examples (AEs) to raise the cost of malicious image editing and copyright violation powered by latent diffusion models (LDMs). Despite their successes, a few have studied the surrogate model they used to generate AEs. In this paper, from the perspective of adversarial transferability, we investigate how the surrogate model's property influences the performance of AEs for LDMs. Specifically, we view the time-step sampling in the Monte-Carlo-based (MC-based) adversarial attack as selecting surrogate models. We find that the smoothness of surrogate models at different time steps differs, and we substantially improve the performance of the MC-based AEs by selecting smoother surrogate models. In the light of the theoretical framework on adversarial transferability in image classification, we also conduct a theoretical analysis to explain why smooth surrogate models can also boost AEs for LDMs.

cs.CV↗

arXiv2022

The Enemy of My Enemy is My Friend: Exploring Inverse Adversaries for Improving Adversarial Training

Although current deep learning techniques have yielded superior performance on various computer vision tasks, yet they are still vulnerable to adversarial examples. Adversarial training and its variants have been shown to be the most effective approaches to defend against adversarial examples. These methods usually regularize the difference between output probabilities for an adversarial and its corresponding natural example. However, it may have a negative impact if the model misclassifies a natural example. To circumvent this issue, we propose a novel adversarial training scheme that encourages the model to produce similar outputs for an adversarial example and its ``inverse adversarial'' counterpart. These samples are generated to maximize the likelihood in the neighborhood of natural examples. Extensive experiments on various vision datasets and architectures demonstrate that our training method achieves state-of-the-art robustness as well as natural accuracy. Furthermore, using a universal version of inverse adversarial examples, we improve the performance of single-step adversarial training techniques at a low computational cost.

cs.CV↗

arXiv2022

Restricted Black-box Adversarial Attack Against DeepFake Face Swapping

DeepFake face swapping presents a significant threat to online security and social media, which can replace the source face in an arbitrary photo/video with the target face of an entirely different person. In order to prevent this fraud, some researchers have begun to study the adversarial methods against DeepFake or face manipulation. However, existing works focus on the white-box setting or the black-box setting driven by abundant queries, which severely limits the practical application of these methods. To tackle this problem, we introduce a practical adversarial attack that does not require any queries to the facial image forgery model. Our method is built on a substitute model persuing for face reconstruction and then transfers adversarial examples from the substitute model directly to inaccessible black-box DeepFake models. Specially, we propose the Transferable Cycle Adversary Generative Adversarial Network (TCA-GAN) to construct the adversarial perturbation for disrupting unknown DeepFake systems. We also present a novel post-regularization module for enhancing the transferability of generated adversarial examples. To comprehensively measure the effectiveness of our approaches, we construct a challenging benchmark of DeepFake adversarial attacks for future development. Extensive experiments impressively show that the proposed adversarial attack method makes the visual quality of DeepFake face images plummet so that they are easier to be detected by humans and algorithms. Moreover, we demonstrate that the proposed algorithm can be generalized to offer face image protection against various face translation methods.

cs.CV↗

arXiv2022

GIAOTracker: A comprehensive framework for MCMOT with global information and optimizing strategies in VisDrone 2021

In recent years, algorithms for multiple object tracking tasks have benefited from great progresses in deep models and video quality. However, in challenging scenarios like drone videos, they still suffer from problems, such as small objects, camera movements and view changes. In this paper, we propose a new multiple object tracker, which employs Global Information And some Optimizing strategies, named GIAOTracker. It consists of three stages, i.e., online tracking, global link and post-processing. Given detections in every frame, the first stage generates reliable tracklets using information of camera motion, object motion and object appearance. Then they are associated into trajectories by exploiting global clues and refined through four post-processing methods. With the effectiveness of the three stages, GIAOTracker achieves state-of-the-art performance on the VisDrone MOT dataset and wins the 3rd place in the VisDrone2021 MOT Challenge.

cs.CV↗

arXiv2021

Oscillatory shear flows of dense suspensions at imposed pressure: Rheology and micro-structure

Oscillatory shear has been widely used to study the rheological properties of suspensions under unsteady shear. Furthermore, recent works have shown that oscillatory flows can improve the flowability of dense suspensions. While most studies have been done under constant volume we here study oscillatory shear flows of a two-dimensional suspensions using a normal pressure-controlled set-up. To characterise the rheology, we introduce both a complex macroscopic friction coefficient $μ^{*}$, following the convention of the complex viscosity $η^{*}$, and a shear-rate averaged viscous number $J'$. The rheology and microstructure of dense suspensions are studied by systematically varying both the strain magnitude $γ_0$ and $J'$ using numerical simulations. We study both suspensions composed of frictional ($μ_p=0.4$) or frictionless ($μ_p=0$) particles and find that the critical values, as $J' \to 0$, of both the complex macroscopic friction and the number of contacts, both total and sliding, decrease with decreasing $γ_0$. For suspensions composed of frictional particles we also find that the critical (i.e.~the shear jamming) packing fraction $ϕ_c$ increase with decreasing $γ_0$. In both cases, frictional and frictionless, we find that the rheological response approaching the shear jamming turns from a viscous to an elastic response as $γ_0$ is lowered below $\sim0.33$.

cond-mat.soft↗

arXiv2020

Extended $μ(J)$-rheology for dense suspensions at oscillatory shear flows

Recent studies have highlighted that oscillatory and time-dependent shear flows might help increase flowability of dense suspensions. While most focus has been on cross-flows we here study a simple two-dimensional suspensions where we apply simultaneously oscillatory and stationary shear along the same direction. We first show that the viscosities in this set-up significantly decrease with an increasing magnitude of the oscillations, contrary to previous claims. The decrease can be attributed to the large decrease in the number of contacts and an altered microstructure as the magnitude of the oscillation is increased. As a sub-result we find both an extension to the $μ(J)$-rheology, a constitutive relationship between the shear stresses and the shear rate, valid for pure oscillatory flows and with a higher shear-jammed packing fraction for suspensions composed of frictional particles compared to steady-shear conditions.

cond-mat.soft↗

arXiv2017

Discontinuous shear thickening of dense suspensions under confining pressure

We use 2D numerical simulations to study dense suspensions of non-Brownian hard particles using the Critical Load Model (CLM) under constant confining pressures. This simple model shows discontinuous shear thickening (DST) as the tangential forces get activated upon increased shear stresses. By parameterizing a simple binary system of frictional and non-frictional particles of different proportions we show that the jamming packing fraction, at which the viscosity diverges, is controlled by the fraction of frictional contacts. The viscosity of dense suspensions can thereby be expressed as a function of the fraction of frictional contacts as well as the packing fraction of solid particles. In addition, we show that there exists a simple relationship between the fraction of frictional contacts and the two control parameters (under confining pressure): the viscous number J and the ratio between the repulsive barrier force and confining pressure. Under confining pressures the viscosity curves are found to depend on the shear protocol, with the possibility of yielding negative dynamic compressibility.

cond-mat.soft↗