Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Evaluating Deep Learning-Based Face Recognition for Infants and Toddlers: Impact of Age Across Developmental Stages

Afzal Hossain, Mst Rumana Sumi, Stephanie Schuckers

TL;DR

This study addresses the challenge of reliable face recognition for infants and toddlers, where rapid facial growth and limited data complicate longitudinal verification. It evaluates four deep-learning FR models (FaceNet, ArcFace, MagFace, CosFace) on the Infants and Toddlers Longitudinal Face (ITLF) dataset collected over seven sessions across 24 months (ages 0–3). The authors show substantial degradation in verification performance over time, with TAR at FAR 0.1% dropping from about 30.7% for 0–6 months to around 64.7% for 2.5–3 years, and they mitigate this drift using a Domain-Adversarial Neural Network (DANN) that improves TAR by more than 12%. The results highlight the need for age-adaptive, privacy-preserving biometric solutions in smart-city contexts and motivate federated or decentralized learning approaches. This work provides a foundational dataset and methodology for future age-aware biometric systems.

Abstract

Face recognition for infants and toddlers presents unique challenges due to rapid facial morphology changes, high inter-class similarity, and limited dataset availability. This study evaluates the performance of four deep learning-based face recognition models FaceNet, ArcFace, MagFace, and CosFace on a newly developed longitudinal dataset collected over a 24 month period in seven sessions involving children aged 0 to 3 years. Our analysis examines recognition accuracy across developmental stages, showing that the True Accept Rate (TAR) is only 30.7% at 0.1% False Accept Rate (FAR) for infants aged 0 to 6 months, due to unstable facial features. Performance improves significantly in older children, reaching 64.7% TAR at 0.1% FAR in the 2.5 to 3 year age group. We also evaluate verification performance over different time intervals, revealing that shorter time gaps result in higher accuracy due to reduced embedding drift. To mitigate this drift, we apply a Domain Adversarial Neural Network (DANN) approach that improves TAR by over 12%, yielding features that are more temporally stable and generalizable. These findings are critical for building biometric systems that function reliably over time in smart city applications such as public healthcare, child safety, and digital identity services. The challenges observed in early age groups highlight the importance of future research on privacy preserving biometric authentication systems that can address temporal variability, particularly in secure and regulated urban environments where child verification is essential.

Evaluating Deep Learning-Based Face Recognition for Infants and Toddlers: Impact of Age Across Developmental Stages

TL;DR

This study addresses the challenge of reliable face recognition for infants and toddlers, where rapid facial growth and limited data complicate longitudinal verification. It evaluates four deep-learning FR models (FaceNet, ArcFace, MagFace, CosFace) on the Infants and Toddlers Longitudinal Face (ITLF) dataset collected over seven sessions across 24 months (ages 0–3). The authors show substantial degradation in verification performance over time, with TAR at FAR 0.1% dropping from about 30.7% for 0–6 months to around 64.7% for 2.5–3 years, and they mitigate this drift using a Domain-Adversarial Neural Network (DANN) that improves TAR by more than 12%. The results highlight the need for age-adaptive, privacy-preserving biometric solutions in smart-city contexts and motivate federated or decentralized learning approaches. This work provides a foundational dataset and methodology for future age-aware biometric systems.

Abstract

Face recognition for infants and toddlers presents unique challenges due to rapid facial morphology changes, high inter-class similarity, and limited dataset availability. This study evaluates the performance of four deep learning-based face recognition models FaceNet, ArcFace, MagFace, and CosFace on a newly developed longitudinal dataset collected over a 24 month period in seven sessions involving children aged 0 to 3 years. Our analysis examines recognition accuracy across developmental stages, showing that the True Accept Rate (TAR) is only 30.7% at 0.1% False Accept Rate (FAR) for infants aged 0 to 6 months, due to unstable facial features. Performance improves significantly in older children, reaching 64.7% TAR at 0.1% FAR in the 2.5 to 3 year age group. We also evaluate verification performance over different time intervals, revealing that shorter time gaps result in higher accuracy due to reduced embedding drift. To mitigate this drift, we apply a Domain Adversarial Neural Network (DANN) approach that improves TAR by over 12%, yielding features that are more temporally stable and generalizable. These findings are critical for building biometric systems that function reliably over time in smart city applications such as public healthcare, child safety, and digital identity services. The challenges observed in early age groups highlight the importance of future research on privacy preserving biometric authentication systems that can address temporal variability, particularly in secure and regulated urban environments where child verification is essential.
Paper Structure (17 sections, 5 figures, 1 table)

This paper contains 17 sections, 5 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Longitudinal Data Collection
  4. Camera Placement and Setup
  5. Camera and Software Settings
  6. Collection Process and Conditions
  7. Methodology
  8. Face Detection and Recognition
  9. Time-Invariant Embedding Learning via Domain Adaptation
  10. Results and Discussion
  11. Longitudinal Evaluation of Face Recognition Performance in Infants, and Toddlers
  12. Performance Evaluation for Intra-Session Face Recognition
  13. Impact of Temporal Aging on Face Recognition Performance
  14. Age-Based Performance Analysis of Face Recognition Models
  15. Time-Invariant Embedding Learning via Domain Adaptation
  16. ...and 2 more sections

Figures (5)

  • Figure 1: Age progression of a subject across different sessions in the Infants and Toddlers Longitudinal Face Image Database.
  • Figure 2: Distribution of collected images by age groups and their respective percentages.
  • Figure 3: True Acceptance Rate (TAR) of the FaceNet, ArcFace, MagFace, and CosFace models at a False Acceptance Rate (FAR) of 0.1% within a session.
  • Figure 4: True Acceptance Rate (TAR) of FaceNet, ArcFace, MagFace, and CosFace at a False Acceptance Rate (FAR) of 0.1%, comparing initial captures with subsequent images collected over varying time intervals.
  • Figure 5: True Acceptance Rate (TAR) at a False Acceptance Rate (FAR) of 0.1% across age groups for 16–20 months and 20–24 months time intervals with and without domain-adapted features using MagFace.