Zero-KnowledgeBiometricIdentityProof

Made in: 11/2025Research paper publishedPurpose: Project for Cryptography and Network Security Subject

A research-backed authentication system designed to securely verify patient identity in healthcare environments without relying on explicit identifiers. The system combines deep biometric representations, cryptographic key binding, and hardware-backed security to enable fast, privacy-preserving, and regulation-compliant access to sensitive clinical data.

Project preview 1
Project preview 2
Project preview 3
Project preview 4

Tech Stack

Python
Python
PostgreSQL
PostgreSQL

ADDITIONAL TOOLS

AES-256 Encryption
HMAC
FAISS
Hardware Security Module (HSM)
Deep Learning
Biometric Fingerprint Embeddings

Features

  • 1

    Identifier-Free Authentication

    Resolves user identity directly through biometric similarity search without using usernames, patient IDs, or personal identifiers.

  • 2

    Biometric–Passcode Key Binding

    Derives cryptographic keys using an HMAC-based construction that securely combines a user passcode with a stable biometric reference.

  • 3

    Hardware-Enforced Key Protection

    Uses Hardware Security Modules (HSMs) to isolate and protect encryption keys, ensuring zero plaintext exposure even under database compromise.

  • 4

    Encrypted Biometric Storage & Search

    Stores biometric embeddings encrypted with AES-256 and performs similarity search using FAISS within protected security boundaries.

  • 5

    High-Performance Authentication

    Achieves low-latency authentication and high throughput on commodity hardware, suitable for real-time clinical workflows.

  • 6

    Regulatory-Compliant Security Design

    Designed to align with GDPR and healthcare security requirements while maintaining confidentiality, integrity, and recoverability.