Low-Complexity CNN-Based Classification of Electroneurographic Signals

TL;DR

This paper presents MobilESCAPE-Net, a lightweight CNN architecture that significantly reduces computational complexity while maintaining high accuracy for real-time electroneurographic signal classification, suitable for implantable devices.

Contribution

The study introduces MobilESCAPE-Net, a novel low-complexity CNN that outperforms existing models in efficiency, enabling real-time ENG signal classification in resource-limited settings.

Findings

Achieves comparable accuracy and F1-score to ESCAPE-Net

Reduces trainable parameters by 99.9%

Lowers floating point operations by 92.47%

Abstract

Peripheral nerve interfaces (PNIs) facilitate neural recording and stimulation for treating nerve injuries, but real-time classification of electroneurographic (ENG) signals remains challenging due to constraints on complexity and latency, particularly in implantable devices. This study introduces MobilESCAPE-Net, a lightweight architecture that reduces computational cost while maintaining and slightly improving classification performance. Compared to the state-of-the-art ESCAPE-Net, MobilESCAPE-Net achieves comparable accuracy and F1-score with significantly lower complexity, reducing trainable parameters by 99.9\% and floating point operations per second by 92.47\%, enabling faster inference and real-time processing. Its efficiency makes it well-suited for low-complexity ENG signal classification in resource-constrained environments such as implantable devices.