Compact Minutiae Descriptor for Fingerprint Matching on Resource-Constrained Microcontrollers
Journal of Engineering System · 2026-09 · 4 (3) · 104-112
DOI: 10.62517/jes.202602310 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Engineering System, 2026-09. Verfasst von Junda Li und 4 weiteren Autor:innen. Die Arbeit nennt 15 Literaturangaben.
Abstract (Originalsprache)
Fingerprint-based biometric authentication is increasingly demanded in IoT and embedded devices, yet state-of-the-art matching algorithms require memory and compute resources that far exceed typical microcontroller capabilities. We present a lightweight fingerprint matching framework designed for ARM Cortex-M class microcontrollers with sub-1 MB SRAM. Our approach introduces four components: (1) a Compact Minutiae Descriptor (CMD) that reduces per-minutia storage from 48 bytes to 14 bytes through angular quantization and local topology encoding; (2) a fixed-point scoring engine using lookup tables that eliminates floating-point dependencies; (3) a hierarchical two-level indexing scheme that prunes 88% of candidates before detailed matching; and (4) a memory-aware enrollment strategy that adaptively adjusts template precision to fit available SRAM. Evaluated via cycle-accurate simulation on 12 FVC benchmark databases (FVC2000, FVC2002, FVC2004) with score distributions calibrated to published competition statistics, CMD achieves a projected mean Equal Error Rate (EER) of 1.95%, comparable to the Minutia Cylinder-Code (MCC, 1.91%) while requiring only 42 KB Flash and 8 KB base SRAM. On a simulated Cortex-M4 at 168 MHz, 1:100 identification completes in 6.9 ms, representing a 68x projected speedup over Bozorth3. These results suggest that competitive fingerprint matching is feasible on resource-constrained platforms without cloud offloading, enabling on-device authentication for edge deployments.
Zitieren
APA
Li, J., Li, W., Jia, J., Fan, S., & Cui, X. (2026). Compact Minutiae Descriptor for Fingerprint Matching on Resource-Constrained Microcontrollers. Journal of Engineering System, 4(3), 104-112. https://doi.org/10.62517/jes.202602310
Vancouver
Li J, Li W, Jia J, Fan S, Cui X. Compact Minutiae Descriptor for Fingerprint Matching on Resource-Constrained Microcontrollers. Journal of Engineering System. 2026;4(3):104-112. doi:10.62517/jes.202602310
BibTeX
@article{Li2026,
author = {Li, Junda and Li, Wen and Jia, Jing and Fan, Suhui and Cui, Xiaoran},
title = {Compact Minutiae Descriptor for Fingerprint Matching on Resource-Constrained Microcontrollers},
journal = {Journal of Engineering System},
year = {2026},
volume = {4},
number = {3},
pages = {104--112},
doi = {10.62517/jes.202602310}
}
EUVIMED