Robust Machine-vision models against adversarial attacks in Industry 4.0

<ns3:p>Background Industrial production systems increasingly depend on deep learning–based machine-vision models for automated inspection, quality assurance, and process monitoring in Industry 4.0 environments. While these models achieve high accuracy under normal conditions, they are vulnerable to adversarial attacks that introduce subtle, visually imperceptible perturbations. Such attacks can lead to defect masking, misclassification, and potentially unsafe system behaviour. Despite extensive research on adversarial robustness in generic vision tasks, systematic studies tailored to industrial machine-vision pipelines remain limited. Methods This work analyses adversarial threats against industrial machine-vision systems and proposes a novel defence mechanism, Consistency-based Machine-Vision Adversarial Defence (C-MVAD). The approach is designed as a drop-in extension for convolutional neural networks and combines adversarial training, lightweight regularisation, and embedding-consistency checks. Robustness is evaluated against three gradient-based attacks, namely Fast Gradient Sign Method (FGSM), Basic Iterative Method (BIM), and Projected Gradient Descent (PGD). Experiments are conducted on an industrial defect detection benchmark using a compact convolutional backbone suitable for edge deployment. Performance is assessed in terms of classification accuracy under attack, adversarial detection capability, and runtime overhead. Results The baseline model exhibits a substantial accuracy degradation under BIM and PGD attacks, with accuracy dropping below 30 % for PGD at moderate perturbation levels. In contrast, C-MVAD maintains significantly higher robustness across all attack types. For PGD attacks, the proposed method achieves an Area Under the Curve (AUC) of 0.995 and True Positive Rates exceeding 86 % at 1 % False Positive Rate. The additional inference latency introduced by C-MVAD is limited to 0.8 ms per image, corresponding to a 21.1 % overhead, remaining well within typical industrial real-time constraints. Conclusions The results demonstrate that industrial machine-vision models are highly susceptible to adversarial manipulation, but that practical and efficient defences are feasible. C-MVAD provides strong adversarial robustness and detection performance with minimal computational overhead, making it suitable for real-time industrial deployment. The proposed consistency-based approach offers a scalable and model-agnostic pathway towards improving the security and reliability of machine-vision systems in Industry 4.0 environments.</ns3:p>