Abstract

This paper investigates the potential of memristor-based neuromorphic chips as a hardware platform for real-time biometric
identification. Memristors, owing to their memory-like resistive switching and biological synapse-like behavior, are promising
components for brain-inspired computing architectures. We explore device characteristics, neuromorphic circuit designs, and
system-level integration for biometric signal recognition, including face, fingerprint, and multimodal biometric patterns.
Challenges such as device variability, limited resistance states, and system integration are discussed, along with future
research directions for scalable and energy-efficient biometric processors. The research demonstrates that memristive
neuromorphic chips can outperform traditional von Neumann systems in speed, energy efficiency, and on-chip learning
capability.