BabyFlow: 3D modeling of realistic and expressive infant faces

2025-12-22

AI Summary

Infant facial data scarcity and highly spontaneous, variable expressions hinder accurate 3D facial modeling and early screening of neurodevelopmental disorders. Method: We propose the first disentangled identity-expression generative framework tailored for infants. (1) Normalizing flows model nonlinear facial expression variations, abandoning restrictive linear assumptions; (2) a cross-age 3D expression transfer strategy mitigates severe annotation scarcity in infant data; (3) a diffusion-based generative architecture is integrated with explicit 3D geometric constraints to enable high-fidelity, synchronized 2D/3D synthesis. Results: Experiments demonstrate substantial improvements in 3D reconstruction accuracy—particularly in dynamic regions (mouth, eyes, nose)—robust identity-preserving expression editing, and generation of geometrically consistent, photorealistic 2D infant faces. The framework establishes an interpretable, controllable 3D analytical foundation for intelligent developmental disorder screening under few-shot conditions.

Technology Category

• Computer Vision: 3D Computer Vision
• Natural Language Processing: Generation
• Machine Learning: Deep Generative Models & Autoencoders

Application Category

• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Web Mining and Content Analysis: Large pretrained models with web data
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

Early detection of developmental disorders can be aided by analyzing infant craniofacial morphology, but modeling infant faces is challenging due to limited data and frequent spontaneous expressions. We introduce BabyFlow, a generative AI model that disentangles facial identity and expression, enabling independent control over both. Using normalizing flows, BabyFlow learns flexible, probabilistic representations that capture the complex, non-linear variability of expressive infant faces without restrictive linear assumptions. To address scarce and uncontrolled expressive data, we perform cross-age expression transfer, adapting expressions from adult 3D scans to enrich infant datasets with realistic and systematic expressive variants. As a result, BabyFlow improves 3D reconstruction accuracy, particularly in highly expressive regions such as the mouth, eyes, and nose, and supports synthesis and modification of infant expressions while preserving identity. Additionally, by integrating with diffusion models, BabyFlow generates high-fidelity 2D infant images with consistent 3D geometry, providing powerful tools for data augmentation and early facial analysis.

Problem

• Generates 3D infant face models with independent identity and expression control
• Transfers adult expressions to enrich infant datasets for better modeling
• Improves reconstruction and synthesis of expressive infant faces for medical analysis

Innovation

• Generative AI disentangles identity and expression
• Normalizing flows capture non-linear facial variability
• Cross-age expression transfer enriches infant datasets