Adversarial-Robust Child Face Verification Using Spiking Neural Networks

Child victim identification increasingly employs facial recognition technology, with recent international operations using face matching to identify victims. However, face recognition systems are vulnerable to adversarial attacks, raising concerns that such techniques could be exploited to evade detection. We investigate whether spiking neural networks (SNNs), known for inherent adversarial robustness in image classification, can provide more resilient face verification for child protection systems. On the YLFW child face dataset, we evaluate SNNs against convolutional neural networks (CNNs), Vision Transformers (ViTs), and state-of-the-art face recognition models under gradient-based adversarial attacks. Our SNN achieves 14.4% equal error rate (EER) under Fast Gradient Sign Method (FGSM) attack (ε=0.10) compared to 60.8% for CNN and 61.7% for ViT, a 4.2× improvement in robustness. Under stronger Auto-PGD (APGD-20) attacks, our compact 455Kparameter SNN (23.2% EER) outperforms pre-trained face verification models trained on 17M images: CosFace (29.7– 37.6% EER, 24–65M params) and LVFace (41.7–46.2% EER, 19–256M params). SNNs thus offer a promising architecture for security-critical child face verification, delivering 1.5–4× greater adversarial robustness than conventional architectures without adversarial training and with only modest impact on clean accuracy.