Exploiting Multi-Object Relationships for Detecting Adversarial Attacks in Complex Scenes

TL;DR

This paper introduces a model-agnostic method for detecting adversarial attacks in complex scenes by leveraging language models to verify object co-occurrence relationships, achieving high accuracy without relying on specific detectors.

Contribution

It proposes a novel, generalizable approach using language models for context consistency checks to detect adversarial attacks in diverse scene images.

Findings

High detection accuracy in complex scenes

Independence from specific object detection models

Effective in practical multi-object scenarios

Abstract

Vision systems that deploy Deep Neural Networks (DNNs) are known to be vulnerable to adversarial examples. Recent research has shown that checking the intrinsic consistencies in the input data is a promising way to detect adversarial attacks (e.g., by checking the object co-occurrence relationships in complex scenes). However, existing approaches are tied to specific models and do not offer generalizability. Motivated by the observation that language descriptions of natural scene images have already captured the object co-occurrence relationships that can be learned by a language model, we develop a novel approach to perform context consistency checks using such language models. The distinguishing aspect of our approach is that it is independent of the deployed object detector and yet offers very high accuracy in terms of detecting adversarial examples in practical scenes with multiple objects.