Detection of Adversarial Attacks in Robotic Perception

Ziad Sharawy; Mohammad Nakshbandi; Sorin Mihai Grigorescu

cs.CVMar 30, 2026

Detection of Adversarial Attacks in Robotic Perception

Authors: Ziad Sharawy, Mohammad Nakshbandi, Sorin Mihai Grigorescu

Organizations: Department of Mechatronics and Robotics, Faculty of Electrical Engineering and Computer Science, Transylvania University of Bras,ov, Romania

Abstract

Deep Neural Networks (DNNs) achieve strong performance in semantic segmentation for robotic perception but remain vulnerable to adversarial attacks, threatening safety-critical applications. While robustness has been studied for image classification, semantic segmentation in robotic contexts requires specialized architectures and detection strategies.

Figures & tables

Panel of Figure 1 — Figure 1: Classifier predictions on clean and adversarial images. (a) ResNet‑18 shows unstable predictions with frequent misclassifications. (b) ResNet‑50 demonstrates robust, consistent performance against adversarial perturbations.

Figure 5: Visualization of FGSM adversarial examples Goodfellow et al. (2015b) on a DeepLabV3+ Chen et al. (2018b) model with ResNet-18 He et al. (2016) Deng et al. (2009) Cordts et al. (2016) , showing clean and adversarial inputs with increasing $\epsilon$ , Paszke et al. (2019) .

Explore similar work

AEGIS: A Semantic GAN and Evidential Learning Frameworkfor Robust Adversarial Detection in Vision Sensors
Jun 25, 2026Maher Boughdiri, Mounira Msahli, Albert BifetAdversarial Example Detection Image Classification
LTCI, Département Informatique et Réseaux, Télécom Paris, Institut Polytechnique de Paris, France
SeFaR: Semantic Feature-aware Robustness Testing of Deep Neural Networks
Aug 10, 2026Nusrat Jahan Mozumder, Divya Gopinath, Corina Pasareanu +1Neural Network Verification Neural Network Robustness
University of Virginia · USA · KBR Inc., NASA Ames +1
Multi-Task Consistency-based Detection of Adversarial Attacks
Aug 7, 2026Cong Chen, Jean-Philippe Monteuuis, Jonathan PetitAdversarial Example Detection
Qualcomm Technologies, Inc. San Diego, CA 92121, USA

Jun 25, 2026J/K move · Enter open · S save

AEGIS: A Semantic GAN and Evidential Learning Frameworkfor Robust Adversarial Detection in Vision Sensors

Maher Boughdiri, Mounira Msahli, Albert Bifet

LTCI, Département Informatique et Réseaux, Télécom Paris, Institut Polytechnique de Paris, France

Adversarial Example Detection Image Classification

Open paper AskPDF arXiv

Deep neural networks (DNNs) have shown outstanding performance in visual recognition tasks within vision sensor networks; however, they are still vulnerable to adversarial manipulations and imperceptible perturbations that can lead to erroneous predictions. To address that, this paper presents AEGIS, a semantic aware and uncertainty guided adversarial detection framework designed for robust image classification in vision sensors pipelines. At its core, a SemantiGAN module functions as a multi class semantic discriminator, identifying and filtering visually inconsistent adversarial inputs before they propagate further in the pipeline. For inputs that pass this stage, a stochastic augmentation process generates test time variations, from which handcrafted instability metrics FlipScore, Prediction Inconsistency, Layerwise Cosine Similarity (early and mid layers), and Entropy are computed. These features are aggregated into a compact five dimensional vector and processed by an Evidential Deep Learning (EDL) classifier, which models output evidence using a Dirichlet distribution to yield both class predictions and calibrated uncertainty estimates. Evaluations on the Tiny ImageNet dataset across six categories clean, FGSM, PGD, patch based, functional, and geometric attacks demonstrate the effectiveness of AEGIS. The proposed framework achieves an AUROC of 92.1%, an AUPRC of 90.2%, and an accuracy of 90.7%, outperforming conventional softmax-based detectors in terms of detection performance, robustness, interpretability, and uncertainty calibration.

Abstract

Explore similar work

AEGIS: A Semantic GAN and Evidential Learning Frameworkfor Robust Adversarial Detection in Vision Sensors

SeFaR: Semantic Feature-aware Robustness Testing of Deep Neural Networks

Multi-Task Consistency-based Detection of Adversarial Attacks