Adversarial Patch

2017/12/27 by Tom B. Brown, T. B. Brown, Brown, Tom B. +9 · 6 voices · 36 citations
Computer Science · #Adversarial Robustness in Machine Learning #Anomaly Detection Techniques and Applications #Digital Media Forensic Detection

paper · pdf · doi:10.48550/arxiv.1712.09665

Abstract

We present a method to create universal, robust, targeted adversarial image patches in the real world. The patches are universal because they can be used to attack any scene, robust because they work under a wide variety of transformations, and targeted because they can cause a classifier to output any target class. These adversarial patches can be printed, added to any scene, photographed, and presented to image classifiers; even when the patches are small, they cause the classifiers to ignore the other items in the scene and report a chosen target class. To reproduce the results from the paper, our code is available at https://github.com/tensorflow/cleverhans/tree/master/examples/adversarial_patch

Cited by

• EGRNet: A Lightweight Semantic Segmentation Network with Edge-Gated Refinement and Adversarial Sensing
• AdvNav: Behavior-Guided Black-Box Adversarial Attacks on Vision-Language Navigation
• Read or Ignore? A Unified Benchmark for Typographic-Attack Robustness and Text Recognition in Vision-Language Models
• Reasoning as a Double-Edged Sword: Architecture and Cross-Stage Robustness in Vision-Language-Action Models
• Revealing Physical-World Semantic Vulnerabilities: Universal Adversarial Patch for Infrared Vision-Language Models
• Signal-based Model Access Risk Analysis for AI System Operations Security
• Auditing Inference-Time Defense Evaluation for Multimodal Large Language Models
• BadWAM: When World-Action Models Dream Right but Act Wrong
• SCAR: Semantic Cardiac Adversarial Representation via Spatiotemporal Manifold Optimization in ECG
• Counterfeit Answers: Adversarial Forgery against OCR-Free Document Visual Question Answering
• The Collapse of Patches
• When Robots Obey the Patch: Universal Transferable Patch Attacks on Vision-Language-Action Models
• Adversarial Patch Attacks on Vision-Based Cargo Occupancy Estimation via Differentiable 3D Simulation
• Robust Physical Adversarial Patches Using Dynamically Optimized Clusters
• Phantom Menace: Exploring and Enhancing the Robustness of VLA Models Against Physical Sensor Attacks
• Thermally Activated Dual-Modal Adversarial Clothing against AI Surveillance Systems
• Filtered-ViT: A Robust Defense Against Multiple Adversarial Patch Attacks
• Learning Fourier shapes to probe the geometric world of deep neural networks
• Physically Real-time Infrared Attack against Optical Flow Estimation Networks
• Towards Trustworthy Embodied Intelligence: A Systems Framework and Graded Trustworthiness Levels
• A Single Set of Adversarial Clothes Breaks Multiple Defense Methods in the Physical World
• NAPPure: Adversarial Purification for Robust Image Classification under Non-Additive Perturbations
• Model-agnostic Adversarial Attack and Defense for Vision-Language-Action Models
• Concept-Based Masking: A Patch-Agnostic Defense Against Adversarial Patch Attacks
• Seeing Isn't Believing: Context-Aware Adversarial Patch Synthesis via Conditional GAN
• Vision Transformers: the threat of realistic adversarial patches
• Localizing Adversarial Attacks To Produces More Imperceptible Noise
• Security of World-Model-Based Embodied AI: A Lifecycle of Threats, Defenses, and Evaluation
• Eva-VLA: Evaluating Vision-Language-Action Models' Robustness Under Real-World Physical Variations
• DiffHash: Text-Guided Targeted Attack via Diffusion Models against Deep Hashing Image Retrieval
• Adversarial Patch Attack for Ship Detection via Localized Augmentation
• IPG: Incremental Patch Generation for Generalized Adversarial Patch Training
• Safety of Embodied Navigation: A Survey
• Benchmarking Adversarial Patch Selection and Location
• Scalable and Precise Patch Robustness Certification for Deep Learning Models with Top-k Predictions
• Radio Adversarial Attacks on EMG-based Gesture Recognition Networks

Discussions

• Adversarial Patch [pdf] [hn, 34 points, 3 comments]
• Adversarial Patch [hn, 7 points, 0 comments]
• Adversarial Patch [lobsters, 2 points, 0 comments]
• Adversarial Patch [pdf] [hn, 2 points, 0 comments]
• Adversarial Patch: images that make classifiers ignore other items in a scene [hn, 1 points, 0 comments]
• Just a patch of a psychedelic sticker can fool image recognition. :O https://arxiv.org/pdf/1712.09665.pdf [bsky, 0 points, 0 comments]

Related

• PatchCleanser: Certifiably Robust Defense against Adversarial Patches for Any Image Classifier
• Generating Adversarial yet Inconspicuous Patches with a Single Image
• PatchZero: Defending against Adversarial Patch Attacks by Detecting and Zeroing the Patch
• PAD: Patch-Agnostic Defense against Adversarial Patch Attacks
• Adversarial Patch Attacks and Defences in Vision-Based Tasks: A Survey