Mitigating Longitudinal Performance Degradation in Child Face Recognition Using Synthetic Data

2026-01-04

arXiv.org

AI Summary

This study addresses the challenge of template drift in child facial recognition caused by rapid nonlinear facial growth, which significantly increases verification error rates over time. To mitigate this, the authors propose a novel data augmentation strategy that, for the first time under an identity-isolation protocol, leverages StyleGAN2-ADA to generate synthetic child faces. This approach is combined with fine-tuning of the MagFace embedding model and incorporates an identity-preserving augmentation scheme alongside a post-generation filtering mechanism to effectively prevent identity leakage and artifact interference. Experimental results demonstrate that, across registration-to-verification intervals of 6 to 36 months, the proposed method substantially reduces error rates, outperforming both pretrained baselines and models fine-tuned exclusively on real data, thereby validating the efficacy of synthetic data in enhancing the robustness of longitudinal child facial recognition.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Life-Long and Continual Learning
• Natural Language Processing: Generation

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Longitudinal face recognition in children remains challenging due to rapid and nonlinear facial growth, which causes template drift and increasing verification errors over time. This work investigates whether synthetic face data can act as a longitudinal stabilizer by improving temporal robustness of child face recognition models. Using an identity disjoint protocol on the Young Face Aging (YFA) dataset, we evaluate three settings: (i) pretrained MagFace embeddings without dataset specific fine-tuning, (ii) MagFace fine-tuned using authentic training faces only, and (iii) MagFace fine-tuned using a combination of authentic and synthetically generated training faces. Synthetic data is generated using StyleGAN2 ADA and incorporated exclusively within the training identities; a post generation filtering step is applied to mitigate identity leakage and remove artifact affected samples. Experimental results across enrollment verification gaps from 6 to 36 months show that synthetic-augmented fine tuning substantially reduces error rates relative to both the pretrained baseline and real only fine tuning. These findings provide a risk aware assessment of synthetic augmentation for improving identity persistence in pediatric face recognition.

Problem

• longitudinal face recognition
• child face recognition
• template drift
• performance degradation
• facial growth

Innovation

• synthetic data augmentation
• longitudinal face recognition
• child face aging
• StyleGAN2-ADA
• identity persistence