Acoustic keyword spotter - optimization from end-user perspective

Acoustic keyword spotter - optimization from end-user perspective

The paper deals with the development of acoustic keyword spotter (KWS) meeting requirements of a real user from the security community. While the basic scheme of the KWS is relatively standard, it uses novel features derived by a hierarchy of neural networks, and score normalization trained to maxim...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Szöke, Igor, Grézl, F, Černocký, J, Fapšo, M, Cipr, Tomáš
Format: Tagungsbericht
Sprache:eng
Schlagworte:
Acoustics
Artificial neural networks
Calibration
Context
Discrete cosine transforms
Estimation
keyword spotting
neural networks
Speech
spoken term detection
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The paper deals with the development of acoustic keyword spotter (KWS) meeting requirements of a real user from the security community. While the basic scheme of the KWS is relatively standard, it uses novel features derived by a hierarchy of neural networks, and score normalization trained to maximize a user-like evaluation metric. The results are reported on a selection of Czech conversational telephone speech (CTS), radio and read data.
DOI:10.1109/SLT.2010.5700849