PulseAugur
EN
LIVE 13:32:15

Fourier Shapes Attack Infrared Object Detection Systems

By PulseAugur Editorial · [1 sources] ·

Researchers have developed a novel method to attack infrared object detection systems by manipulating thermal signatures with geometric shapes. Their approach uses learnable Fourier coefficients to define shape boundaries, which are then mapped to pixel masks using the winding number theorem. This allows for efficient, gradient-based optimization to create shapes that can evade detection, achieving over an 88% attack success rate at distances greater than 25 meters in physical experiments. AI

IMPACT This research demonstrates a new vulnerability in infrared object detection, potentially impacting autonomous driving and surveillance systems.

RANK_REASON The cluster contains a research paper detailing a new method for attacking infrared object detection systems. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Fourier Shapes Attack Infrared Object Detection Systems

COVERAGE [1]

  1. arXiv cs.CV TIER_1 English(EN) · Fan Li ·

    Unleashing the Representational Power of Fourier Shapes for Attacking Infrared Object Detection

    Infrared object detection is crucial for perception in autonomous driving and surveillance but remains vulnerable to physical adversarial attacks. Unlike in the RGB domain, where attacks rely on color texture, infrared attacks must manipulate thermal signatures, making the geomet…

RELATED ENTITIES

RELATED TOPICS