SafeTriage: Facial Video De-identification for Privacy-Preserving Stroke Triage

2025-06-19

AI Summary

Balancing facial video privacy protection and diagnostic utility in stroke emergency triage remains challenging. Method: This paper proposes a lightweight de-identification framework that decouples identity and motion features via a pre-trained Video Motion Transfer (VMT) model, and adapts to patient-specific motion distributions using conditional generative visual prompting—without fine-tuning the backbone network. Results: The synthesized videos preserve critical stroke-related facial motion patterns while reducing re-identification attack success rates to <1.5%. Clinical expert evaluation and AUC analysis confirm high diagnostic consistency (ΔAUC < 0.02). To our knowledge, this is the first work to integrate motion transfer with conditional visual prompting for medical video de-identification, simultaneously ensuring compliance with GDPR/HIPAA regulations and maintaining reliability for AI-assisted stroke diagnosis.

Technology Category

• Computer Vision: Motion & Tracking
• Machine Learning: Privacy
• Natural Language Processing: Safety and Robustness

Application Category

• Security and Privacy: Data transparency and provenance
• Responsible Web: Consent frameworks and practices on the web
• User Modeling, Personalization and Recommendation: Accountability, Transparency, and Ethics for personalization

Abstract

Effective stroke triage in emergency settings often relies on clinicians'ability to identify subtle abnormalities in facial muscle coordination. While recent AI models have shown promise in detecting such patterns from patient facial videos, their reliance on real patient data raises significant ethical and privacy challenges -- especially when training robust and generalizable models across institutions. To address these concerns, we propose SafeTriage, a novel method designed to de-identify patient facial videos while preserving essential motion cues crucial for stroke diagnosis. SafeTriage leverages a pretrained video motion transfer (VMT) model to map the motion characteristics of real patient faces onto synthetic identities. This approach retains diagnostically relevant facial dynamics without revealing the patients'identities. To mitigate the distribution shift between normal population pre-training videos and patient population test videos, we introduce a conditional generative model for visual prompt tuning, which adapts the input space of the VMT model to ensure accurate motion transfer without needing to fine-tune the VMT model backbone. Comprehensive evaluation, including quantitative metrics and clinical expert assessments, demonstrates that SafeTriage-produced synthetic videos effectively preserve stroke-relevant facial patterns, enabling reliable AI-based triage. Our evaluations also show that SafeTriage provides robust privacy protection while maintaining diagnostic accuracy, offering a secure and ethically sound foundation for data sharing and AI-driven clinical analysis in neurological disorders.

Problem

• De-identify patient facial videos for privacy
• Preserve stroke-relevant facial motion cues
• Enable secure AI-based stroke triage

Innovation

• Uses pretrained video motion transfer model
• Introduces conditional generative model for adaptation
• Preserves stroke-relevant facial patterns effectively