This paper introduces a groundbreaking paradigm shift in biometric cryptography by proposing the elimination of stored "Helper Data" in Fuzzy Extractors. Traditional biometric cryptosystems, such as FIDO2 or HSM-based solutions, inherently rely on storing Helper Data (P) in non-volatile memory to correct biometric read errors. We argue that this storage requirement compromises the "Zero-Storage" (Stateless) security model, as the Helper Data itself acts as a static footprint of the user's identity. We propose the HSKG (Hierarchical Spatiotemporal Key Generation) protocol, which mathematically substitutes stored Helper Data with real-time environmental and biological observations. By defining Helper Data as a function of deterministic spatiotemporal metrics (GPS, Nano-second Timestamp) and biological liveness entropy (Heart Rate Variability), we prove that the error-correcting information can be derived from the user's immediate reality rather than a database. The cryptographic key is synthesized on-the-fly and vaporized in 0.024 seconds, achieving Information-Theoretic Security where the key reconstruction is physically impossible without the specific convergence of the user's location, time, and vital signs.
Min Ho Jung (Wed,) studied this question.