English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Object detection has found extensive applications in various tasks, but it isalso susceptible to adversarial patch attacks. Existing defense methods oftennecessitate modifications to the target model or result in unacceptable timeoverhead. In this paper, we adopt a counterattack approach, following theprinciple of "fight fire with fire," and propose a novel and generalmethodology for defending adversarial attacks. We utilize an active defensestrategy by injecting two types of defensive patches, canary and woodpecker,into the input to proactively probe or weaken potential adversarial patcheswithout altering the target model. Moreover, inspired by randomizationtechniques employed in software security, we employ randomized canary andwoodpecker injection patterns to defend against defense-aware attacks. Theeffectiveness and practicality of the proposed method are demonstrated throughcomprehensive experiments. The results illustrate that canary and woodpeckerachieve high performance, even when confronted with unknown attack methods,while incurring limited time overhead. Furthermore, our method also exhibitssufficient robustness against defense-aware attacks, as evidenced by adaptiveattack experiments.

Fight Fire with Fire: Combating Adversarial Patch Attacks using Pattern-randomized Defensive Patches