Hybrid Spiking Convolutional Neural Network (H-SCNN) on AudioMNIST

Authors

• Ludvig B. Oliver, Air Force Institute of Technology
• Kenneth M. Hopkinson, Air Force Institute of TechnologyFollow

Document Type

Article

Publication Date

3-24-2026

Abstract

Excerpt: Radio frequency fingerprinting (RFF) aims to recognize transmitters based on unique RF characteristics from naturally embedded hardware imperfections. Common approaches often use artificial neural networks (ANNs) to accomplish this task, showing promise for physical-layer security; however, neuromorphic implementations are minimal. Merging these two efforts, this paper presents a hybrid spiking convolutional neural network (HSCNN) model that classifies the audioMNIST dataset, serving as a proxy for radio frequencies (RF) due to access limitations.

Comments

© 2026 IEEE.

The full paper is available from the publisher via subscription or purchase using the DOI link below.

DOI

10.1109/SATC69565.2026.11542330

Source Publication

2026 IEEE 2nd International Conference on Secure IoT, Assured and Trusted Computing (SATC)

Recommended Citation

L. B. Oliver and K. M. Hopkinson, "Hybrid Spiking Convolutional Neural Network (H-SCNN) on AudioMNIST," 2026 IEEE 2nd International Conference on Secure IoT, Assured and Trusted Computing (SATC), Houston, TX, USA, 2026, pp. 1-5, doi: 10.1109/SATC69565.2026.11542330.