Adversarial Examples for Semantic Segmentation and Object Detection

Xie, Cihang; Wang, Yunlong; Zhang, Zhishuai; Zhou, Yuyin; Xie, Lingxi; Yuille, Alan

doi:10.1109/iccv.2017.153

Cited by 721 publications

(613 citation statements)

References 29 publications

Supporting

Mentioning

605

Contrasting

Unclassified

Order By: Relevance

“…Many different attack methods for object detectors have been developed very recently [57,32,6,11,55,23,22,31]. Although there are many differences in the formulations of these attacks, when viewed from the multi-task learning…”

Section: Detection Attacks Guided By Task Lossesmentioning

confidence: 99%

“…Components loss cls loss loc T N T N ShapeShifter [6] DFool [32], PhyAttack [11] DAG [57], Transfer [55] DPatch [31] RAP [23] BPatch [22] Table 1. Analysis of existing attack methods for object detection.…”

Section: Attacks For Object Detectionmentioning

confidence: 99%

“…From this perspective, we can categorize existing attack methods as in Table 1. It is clear that some methods use classification loss [6,32,11,57,55] while other methods also incorporated localization loss [31,23,22]. There are two perspectives for explaining the effectiveness of individual task loss in generating attacks: i) the classification and localization tasks share a common base-net, implying that the weakness in the base-net will be shared among all tasks built upon it; ii) while the classification and localization outputs have dedicated branches for each task beyond the shared base-net, they are coupled in the testing phase due to the usage of NMS, which jointly use class scores and bounding box locations for redundant prediction pruning.…”

Section: Attacks For Object Detectionmentioning

confidence: 99%

“…This is one of the reason why adversarial images generated based on individual task losses (e.g. classification loss [57]) can effectively attack object detectors.…”

Section: The Roles Of Task Losses In Robustnessmentioning

confidence: 99%

See 3 more Smart Citations

Towards Adversarially Robust Object Detection

Zhang¹,

Wang²

2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

Self Cite

104

105

View full text Add to dashboard Cite

Object detection is an important vision task and has emerged as an indispensable component in many vision system, rendering its robustness as an increasingly important performance factor for practical applications. While object detection models have been demonstrated to be vulnerable against adversarial attacks by many recent works, very few efforts have been devoted to improving their robustness. In this work, we take an initial attempt towards this direction.We first revisit and systematically analyze object detectors and many recently developed attacks from the perspective of model robustness. We then present a multi-task learning perspective of object detection and identify an asymmetric role of task losses. We further develop an adversarial training approach which can leverage the multiple sources of attacks for improving the robustness of detection models. Extensive experiments on PASCAL-VOC and MS-COCO verified the effectiveness of the proposed approach.

show abstract

Section: Detection Attacks Guided By Task Lossesmentioning

confidence: 99%

Section: Attacks For Object Detectionmentioning

confidence: 99%

Section: Attacks For Object Detectionmentioning

confidence: 99%

“…This is one of the reason why adversarial images generated based on individual task losses (e.g. classification loss [57]) can effectively attack object detectors.…”

Section: The Roles Of Task Losses In Robustnessmentioning

confidence: 99%

See 2 more Smart Citations

Towards Adversarially Robust Object Detection

Zhang¹,

Wang²

2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

Self Cite

104

105

View full text Add to dashboard Cite

show abstract

“…pedestrians. In contrast, Xie et al [19] proposed an approach to generate adversarial examples for image semantic segmentation and object detection without attacking targets. However, a random segmentation result should be specified so that the adversaries can be inferred.…”

Section: Related Workmentioning

confidence: 99%

Intelligent Image Synthesis to Attack a Segmentation CNN Using Adversarial Learning

Chen

Bentley

Mori

et al. 2019

Simulation and Synthesis in Medical Imaging

View full text Add to dashboard Cite

Deep learning approaches based on convolutional neural networks (CNNs) have been successful in solving a number of problems in medical imaging, including image segmentation. In recent years, it has been shown that CNNs are vulnerable to attacks in which the input image is perturbed by relatively small amounts of noise so that the CNN is no longer able to perform a segmentation of the perturbed image with sufficient accuracy. Therefore, exploring methods on how to attack CNN-based models as well as how to defend models against attacks have become a popular topic as this also provides insights into the performance and generalization abilities of CNNs. However, most of the existing work assumes unrealistic attack models, i.e. the resulting attacks were specified in advance. In this paper, we propose a novel approach for generating adversarial examples to attack CNN-based segmentation models for medical images. Our approach has three key features: 1) The generated adversarial examples exhibit anatomical variations (in form of deformations) as well as appearance perturbations; 2) The adversarial examples attack segmentation models so that the Dice scores decrease by a pre-specified amount; 3) The attack is not required to be specified beforehand. We have evaluated our approach on CNN-based approaches for the multi-organ segmentation problem in 2D CT images. We show that the proposed approach can be used to attack different CNN-based segmentation models.

show abstract

A review of research on co‐training

Ning

Wang

et al. 2021

Concurrency and Computation

View full text Add to dashboard Cite

Co-training algorithm is one of the main methods of semi-supervised learning in machine learning, which explores the effective information in unlabeled data by multi-learner collaboration. Based on the development of co-training algorithm, the research work in recent years was further summarized in this article. In particular, three main steps of relevant co-training algorithms are introduced: view acquisition, learners' differentiation, and label confidence estimation. Finally, we summarized the problems existing in the current co-training methods, gave some suggestions for improvement, and looked forward to the future development direction of the co-training algorithm.

show abstract

Adversarial Examples for Semantic Segmentation and Object Detection

Cited by 721 publications

References 29 publications

Towards Adversarially Robust Object Detection

Towards Adversarially Robust Object Detection

Intelligent Image Synthesis to Attack a Segmentation CNN Using Adversarial Learning

A review of research on co‐training

Contact Info

Product

Resources

About