VoteX: Secure Voting System Using Biometric Authentication with CNN-Based Fingerprint Liveness Detection

VoteX demonstrates the practical feasibility of embedding deep-learning-based Presentation Attack Detection into a physical biometric electronic voting workflow. By combining an ESP32 edge terminal, an R307S optical sensor, a Python Flask backend, and a fine-tuned EfficientNetB0 classifier, the system introduces an active liveness gate that screens for artificial replicas before template matching and SQLite database queries can occur.  The system was evaluated across two distinct operational settings. On the 220-sample GreenBit test set provided by the LivDet organizers, the classifier achieved 85.45% accuracy, 80.95% F1-score, 0.9256 ROC-AUC, and 14.64% ACER at an operating threshold of 0.31. When deployed on physical hardware at a calibrated operating cutoff of 0.60 (calibrated on those same 16 scans as a preliminary feasibility check), the pipeline achieved 93.75% accuracy and an ACER of 5.00% across 16 sensor-level fingerprint scans, with 0 of 6 PVA glue spoof presentations accepted. These preliminary results suggest the feasibility of a decoupled client-server architecture for supporting presentation attack detection on commercial hardware without altering existing voter registry mechanisms, though larger independent validation remains required. Duplicate voting is prevented separately via database status checks.   Future research will focus on porting quantized liveness models directly onto edge microcontrollers using TensorFlow Lite Micro to eliminate network dependencies, integrating multi-modal biometric fusion to mitigate ridge-abrasion false rejections, and adopting cryptographic zero-knowledge proofs to decouple voter authentication from ballot tabulation for privacy preservation.

[1]  A. K. Jain, A. Ross, and S. Prabhakar, “An introduction to biometric recognition,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 14, no. 1, pp. 4–20, 2004. [2]    D. Maltoni, D. Maio, A. K. Jain, and S. Prabhakar, Handbook of Fingerprint Recognition, 2nd ed. London, U.K.: Springer, 2009. [3]   International Organization for Standardization, ISO/IEC 30107-3: Information Technology—Biometric Presentation Attack Detection—Part 3: Testing and Reporting, 2nd ed., ISO/IEC, 2023. [4]   M. Tan and Q. V. Le, “EfficientNet: Rethinking model scaling for convolutional neural networks,” in Proceedings of the 36th International Conference on Machine Learning, vol. 97, pp. 6105–6114, 2019. [5]  V. Mura, G. Orrù, R. Casula, A. Sibiriu, G. Loi, P. Tuveri, L. Ghiani, and G. L. Marcialis, “LivDet 2017 Fingerprint Liveness Detection Competition 2017,” in Proc. 11th IAPR Int. Conf. on Biometrics (ICB), 2018. [6]   A. K. Jain, K. Nandakumar, and A. Nagar, “Biometric template security,” EURASIP Journal on Advances in Signal Processing, vol. 2008, p. 579416, 2008. [7]  G. Orrù, R. Casula, P. Tuveri, C. Bazzoni, G. Dessalvi, M. Micheletto, L. Ghiani, and G. L. Marcialis, “LivDet in Action—Fingerprint Liveness Detection Competition 2019,” in Proc. Int. Conf. on Biometrics (ICB), 2019. [8]    Espressif Systems, ESP32 Series Datasheet, Espressif Systems, 2023. [9]   Adafruit Industries, Adafruit Fingerprint Sensor Library Documentation, 2022. [10]     Pallets, Flask Documentation (v3.x), Python Web Framework Documentation, 2023.