Hardware–Software Co-Design of an Audio Feature Extraction Pipeline for Machine Learning Applications
Electronics · 24.02.2024 · 13 (5) · 875
DOI: 10.3390/electronics13050875 ↗Kurz gefasst
Fachartikel in der Zeitschrift Electronics, online erschienen am 24.02.2024. Verfasst von Jure Vreča und 2 weiteren Autor:innen. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 14 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Slovenian Research and Innovation Agency, ECSEL Joint Undertaking.
Abstract (Originalsprache)
Keyword spotting is an important part of modern speech recognition pipelines. Typical contemporary keyword-spotting systems are based on Mel-Frequency Cepstral Coefficient (MFCC) audio features, which are relatively complex to compute. Considering the always-on nature of many keyword-spotting systems, it is prudent to optimize this part of the detection pipeline. We explore the simplifications of the MFCC audio features and derive a simplified version that can be more easily used in embedded applications. Additionally, we implement a hardware generator that generates an appropriate hardware pipeline for the simplified audio feature extraction. Using Chisel4ml framework, we integrate hardware generators into Python-based Keras framework, which facilitates the training process of the machine learning models using our simplified audio features.
Förderung
- Slovenian Research and Innovation Agency
- ECSEL Joint Undertaking
Zitieren
APA
Vreča, J., Pilipović, R., & Biasizzo, A. (2024). Hardware–Software Co-Design of an Audio Feature Extraction Pipeline for Machine Learning Applications. Electronics, 13(5), 875. https://doi.org/10.3390/electronics13050875
Vancouver
Vreča J, Pilipović R, Biasizzo A. Hardware–Software Co-Design of an Audio Feature Extraction Pipeline for Machine Learning Applications. Electronics. 2024;13(5):875. doi:10.3390/electronics13050875
BibTeX
@article{Vrea2024,
author = {Vreča, Jure and Pilipović, Ratko and Biasizzo, Anton},
title = {Hardware–Software Co-Design of an Audio Feature Extraction Pipeline for Machine Learning Applications},
journal = {Electronics},
year = {2024},
volume = {13},
number = {5},
pages = {875},
doi = {10.3390/electronics13050875}
}
EUVIMED