Segment and Complete: Defending Object Detectors against Adversarial Patch Attacks with Robust Patch Detection

TL;DR

This paper introduces SAC, a robust framework that detects and removes adversarial patches in object detection, significantly improving resilience against physical patch attacks without sacrificing accuracy on clean images.

Contribution

The paper proposes a novel detection and removal framework for adversarial patches, including a patch segmenter, self adversarial training, and shape completion, with new dataset annotations.

Findings

SAC achieves superior robustness against adaptive patch attacks.

SAC maintains performance on clean images.

SAC generalizes well to unseen attack types.

Abstract

Object detection plays a key role in many security-critical systems. Adversarial patch attacks, which are easy to implement in the physical world, pose a serious threat to state-of-the-art object detectors. Developing reliable defenses for object detectors against patch attacks is critical but severely understudied. In this paper, we propose Segment and Complete defense (SAC), a general framework for defending object detectors against patch attacks through detection and removal of adversarial patches. We first train a patch segmenter that outputs patch masks which provide pixel-level localization of adversarial patches. We then propose a self adversarial training algorithm to robustify the patch segmenter. In addition, we design a robust shape completion algorithm, which is guaranteed to remove the entire patch from the images if the outputs of the patch segmenter are within a certain Hamming distance of the ground-truth patch masks. Our experiments on COCO and xView datasets demonstrate that SAC achieves superior robustness even under strong adaptive attacks with no reduction in performance on clean images, and generalizes well to unseen patch shapes, attack budgets, and unseen attack methods. Furthermore, we present the APRICOT-Mask dataset, which augments the APRICOT dataset with pixel-level annotations of adversarial patches. We show SAC can significantly reduce the targeted attack success rate of physical patch attacks. Our code is available at https://github.com/joellliu/SegmentAndComplete.