Sparse Adversarial Attack to Object Detection

2020-12-26 · Jiayu Bao

Adversarial examples have gained tons of attention in recent years. Many adversarial attacks have been proposed to attack image classifiers, but few work shift attention to object detectors. In this paper, we propose Sparse Adversarial Attack (SAA) which enables adversaries to perform effective evasion attack on detectors with bounded \emph{l$_{0}$} norm perturbation. We select the fragile position of the image and designed evasion loss function for the task. Experiment results on YOLOv4 and FasterRCNN reveal the effectiveness of our method. In addition, our SAA shows great transferability across different detectors in the black-box attack setting. Codes are available at \emph{https://github.com/THUrssq/Tianchi04}.

📄 PDF Abstract BibTeX arXiv:2012.13692

Code (1)

THUrssq/Tianchi04 공식 구현 pytorch

Tasks

Adversarial AttackObjectobject-detection Object DetectionPosition

Methods 이 논문이 사용한 방법론

Grid Sensitive Grid Sensitive is a trick for object detection introduced by YOLOv4. When we decode the coordinate of the bounding box center $x$…

FPN 설명 없음

Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…

Residual Connection 설명 없음

1x1 Convolution A 1 x 1 Convolution is a convolution with some special properties in that it can be used for dimensionality reduction,…

Average Pooling 설명 없음

Global Average Pooling Global Average Pooling is a pooling operation designed to replace fully connected layers in classical CNNs. The idea is to generate one feature map for each corresponding…

k-Means Clustering k-Means Clustering is a clustering algorithm that divides a training set into $k$ different clusters of examples that are near each other. It works by initializing $k$…

Focused Adversarial Attacks

2022-05-19 · Thomas Cilloni, Charles Walter, Charles Fleming

Recent advances in machine learning show that neural models are vulnerable to minimally perturbed inputs, or adversarial examples. Adversarial algorithms are optimization problems that minimize the accuracy of ML models …

object-detection Object Detection

Sparse Autoencoders as Plug-and-Play Firewalls for Adversarial Attack Detection in VLMs

2026-05-08 · Hao Wang, Yiqun Sun, Pengfei Wei, Lawrence B. Hsieh 외 arxiv

Vision-language models (VLMs) have advanced rapidly and are increasingly deployed in real-world applications, especially with the rise of agent-based systems. However, their safety has received relatively limited attenti…

Domain Generalization Adversarial Attack

Attacking Important Pixels for Anchor-free Detectors

2023-01-26 · Yunxu Xie, Shu Hu, Xin Wang, Quanyu Liao 외

Deep neural networks have been demonstrated to be vulnerable to adversarial attacks: subtle perturbation can completely change the prediction result. Existing adversarial attacks on object detection focus on attacking an…

Adversarial Attack object-detection Object Detection Pose Estimation

Adversarial Attacks against Deep Saliency Models

2019-04-02 · Zhaohui Che, Ali Borji, Guangtao Zhai, Suiyi Ling 외

Currently, a plethora of saliency models based on deep neural networks have led great breakthroughs in many complex high-level vision tasks (e.g. scene description, object detection). The robustness of these models, howe…

Adversarial Attack object-detection Object Detection

To Make Yourself Invisible with Adversarial Semantic Contours

2023-03-01 · Yichi Zhang, Zijian Zhu, Hang Su, Jun Zhu 외

Modern object detectors are vulnerable to adversarial examples, which may bring risks to real-world applications. The sparse attack is an important task which, compared with the popular adversarial perturbation on the wh…

Autonomous Driving Object object-detection Object Detection