Multimodal Biometric Cryptography
Fused physiological features to generate cryptographic keys.
- year
- March 2024 — May 2024
- role
- Course Project
- location
- Amrita Vishwa Vidyapeetham
- category
- Academic
[ flow analysis // multimodal encryption & decryption architecture ]
Context
Engineered a multimodal biometric cryptography architecture that integrates physiological traits to eliminate the security vulnerabilities inherent in single-source authenticators.
Pipeline
Designed a parallel image processing pipeline that simultaneously executes discrete data preprocessing and feature extraction on human palmprints and a 5-finger minutiae array.
Fusion
Implemented a multi-tier feature-level fusion matrix that seamlessly combines morphological palm structures and multi-finger ridge patterns into a single, high-entropy template.
Cryptography
Leveraged the SHA-256 hashing protocol to dynamically convert the fused physiological matrices into unique, bio-inspired cryptographic keys for symmetric data encryption and decryption.
Performance
Achieved a verified 97% authentication accuracy rate during system validation, optimizing the critical equilibrium between cryptographic complexity and matching precision.
next entry
Ericsson Customer Acceptance — Platform R&D