Trustworthy Biometrics and Digital Identity

How do we create, recognize, and protect human identity in a world increasingly populated by synthetic people, avatars, and AI-generated media?

Our longest and most cited line of work spans face, body, and gait recognition; open-set and large-scale identification; long-range and aerial recognition; deepfake detection; and synthetic identity. As digital entities and AI-generated media proliferate, we study how to recognize real people robustly, generate identity-aware synthetic data, detect manipulation, and provision secure identity — with fairness, privacy, and explainability built in.

What we work on

  • Face, body, gait, and multimodal biometrics
  • Open-set and large-scale recognition
  • Long-range and aerial person recognition
  • Robustness across clothing, pose, viewpoint, and time
  • Synthetic biometric data
  • Deepfake detection and media authenticity
  • Privacy-preserving recognition
  • Digital and virtual identity
  • Fairness, explainability, and uncertainty
  • Secure identity for AI-generated digital entities

Selected work

Where it applies

Identity verification Fraud prevention Content authenticity & provenance Public safety & defense Synthetic-data generation Privacy-preserving analytics Avatar & virtual-world identity
← All research