The Vulnerability of Semantic Segmentation Networks to Adversarial Attacks in Autonomous Driving: Enhancing Extensive Environment Sensing

Bär, Andreas; Löhdefink, Jonas; Kapoor, Nikhil; Varghese, Serin; Hüger, Fabian; Schlicht, Peter; Fingscheidt, Tim

doi:10.1109/msp.2020.2983666

Cited by 30 publications

(10 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Segmentation neural networks are not immune to these types of attack. For example, in [ 72 ], it was proven that an adversative attack can fool ICNet [ 73 ], which provides semantic segmentation by controlling an autonomously driving car. The changes in the adversative algorithm applied to the input image are so subtle that a real-world light imperfection or a camera sensor not working properly can recreate the adversative pattern, leading to an accident.…”

Section: Proposed Methodsmentioning

confidence: 99%

NeuronAlg: An Innovative Neuronal Computational Model for Immunofluorescence Image Segmentation

Giacopelli

Migliore

Tegolo

2023

Sensors

View full text Add to dashboard Cite

Background: Image analysis applications in digital pathology include various methods for segmenting regions of interest. Their identification is one of the most complex steps and therefore of great interest for the study of robust methods that do not necessarily rely on a machine learning (ML) approach. Method: A fully automatic and optimized segmentation process for different datasets is a prerequisite for classifying and diagnosing indirect immunofluorescence (IIF) raw data. This study describes a deterministic computational neuroscience approach for identifying cells and nuclei. It is very different from the conventional neural network approaches but has an equivalent quantitative and qualitative performance, and it is also robust against adversative noise. The method is robust, based on formally correct functions, and does not suffer from having to be tuned on specific data sets. Results: This work demonstrates the robustness of the method against variability of parameters, such as image size, mode, and signal-to-noise ratio. We validated the method on three datasets (Neuroblastoma, NucleusSegData, and ISBI 2009 Dataset) using images annotated by independent medical doctors. Conclusions: The definition of deterministic and formally correct methods, from a functional and structural point of view, guarantees the achievement of optimized and functionally correct results. The excellent performance of our deterministic method (NeuronalAlg) in segmenting cells and nuclei from fluorescence images was measured with quantitative indicators and compared with those achieved by three published ML approaches.

show abstract

Section: Proposed Methodsmentioning

confidence: 99%

NeuronAlg: An Innovative Neuronal Computational Model for Immunofluorescence Image Segmentation

Giacopelli

Migliore

Tegolo

2023

Sensors

View full text Add to dashboard Cite

show abstract

“…6G likely enhances specialized techniques for detecting AI-empowered technologies' weaknesses, e.g., AI model assessment, API for scanning vulnerabilities in AI-empowered services. More detail of adversarial attacks and defenses in deep learning can be found in [229]- [232]. In conclusion, the field of AI security protection keeps capturing the interests of academic and industrial societies.…”

Section: Artificial Intelligence's Impact On 6g Securitymentioning

confidence: 99%

Security and privacy for 6G: A survey on prospective technologies and challenges

Nguyen,

Lin,

Cheng

et al. 2021

Preprint

View full text Add to dashboard Cite

Sixth-generation (6G) mobile networks will have to cope with diverse threats on a space-air-ground integrated network environment, novel technologies, and an accessible user information explosion. However, for now, security and privacy issues for 6G remain largely in concept. This survey provides a systematic overview of security and privacy issues based on prospective technologies for 6G in the physical, connection, and service layers, as well as through lessons learned from the failures of existing security architectures and state-of-the-art defenses.Two key lessons learned are as follows. First, other than inheriting vulnerabilities from the previous generations, 6G has new threat vectors from new radio technologies, such as the exposed location of radio stripes in ultra-massive MIMO systems at Terahertz bands and attacks against pervasive intelligence. Second, physical layer protection, deep network slicing, quantum-safe communications, artificial intelligence (AI) security, platform-agnostic security, real-time adaptive security, and novel data protection mechanisms such as distributed ledgers and differential privacy are the top promising techniques to mitigate the attack magnitude and personal data breaches substantially.

show abstract

“…As an extension of classification, semantic segmentation also suffers from adversarial examples [50,2]. Segmentation models applied in real-world safety-critical applications also face potential threats, e.g., in self-driving systems [32,25,19,33,4,36] and in medical image analysis [10,35,12,30]. Hence, the adversarial robustness of segmentation has also raised great attention recently [50,2,51,49,20,42,27,24,49,8,38,53].…”

Section: Introductionmentioning

confidence: 99%

SegPGD: An Effective and Efficient Adversarial Attack for Evaluating and Boosting Segmentation Robustness

Gu¹,

Zhao²,

Tresp³

et al. 2022

Preprint

View full text Add to dashboard Cite

Deep neural network-based image classifications are vulnerable to adversarial perturbations. The image classifications can be easily fooled by adding artificial small and imperceptible perturbations to input images. As one of the most effective defense strategies, adversarial training was proposed to address the vulnerability of classification models, where the adversarial examples are created and injected into training data during training. The attack and defense of classification models have been intensively studied in past years. Semantic segmentation, as an extension of classifications, has also received great attention recently. Recent work shows a large number of attack iterations are required to create effective adversarial examples to fool segmentation models. The observation makes both robustness evaluation and adversarial training on segmentation models challenging. In this work, we propose an effective and efficient segmentation attack method, dubbed SegPGD. Besides, we provide a convergence analysis to show the proposed SegPGD can create more effective adversarial examples than PGD under the same number of attack iterations. Furthermore, we propose to apply our SegPGD as the underlying attack method for segmentation adversarial training. Since SegPGD can create more effective adversarial examples, the adversarial training with our SegPGD can boost the robustness of segmentation models. Our proposals are also verified with experiments on popular Segmentation model architectures and standard segmentation datasets.

show abstract

The Vulnerability of Semantic Segmentation Networks to Adversarial Attacks in Autonomous Driving: Enhancing Extensive Environment Sensing

Cited by 30 publications

References 22 publications

NeuronAlg: An Innovative Neuronal Computational Model for Immunofluorescence Image Segmentation

NeuronAlg: An Innovative Neuronal Computational Model for Immunofluorescence Image Segmentation

Security and privacy for 6G: A survey on prospective technologies and challenges

SegPGD: An Effective and Efficient Adversarial Attack for Evaluating and Boosting Segmentation Robustness

Contact Info

Product

Resources

About