Event-Driven Continuous-Time Feature Extraction for Ultra Low-Power Audio Keyword Spotting
Résumé
In the context of autonomous keyword spotting and sound detection, this paper proposes a low power feature extraction unit generating spectrograms that represent a unique signature allowing the classification of audio signals. This system is composed of a continuous-Time digital signal processing feature extractor combined with a convolutional neural network engine. The study evaluates the hardware requirements to implement the feature extraction unit using an advanced CMOS process. Furthermore, a simulation of the complete system using Matlab® reveals that the recognition accuracy remains higher than 90% while offering a power consumption 4000X lower than a conventional discrete time system.
Mots clés
Audio acoustics
Continuous time systems
Convolutional neural networks
Digital control systems
Digital signal processing
Discrete time control systems
Extraction
MATLAB
Complete system
Continuous-time digital signal processing
Discrete - time systems
Feature extractor
Keyword spotting
Recognition accuracy
Sound detection
Ultra low power
Feature extraction
Domaines
Sciences de l'ingénieur [physics]
Fichier principal
Event-Driven Continuous-Time Feature Extraction for Ultra Low-Power Audio Keyword Spotting_accepted.pdf (569.65 Ko)
Télécharger le fichier
Origine | Fichiers produits par l'(les) auteur(s) |
---|