TU Darmstadt / ULB / TUbiblio

Hierarchical Spectro-Temporal Features for Robust Speech Recognition

Domont, Xavier ; Heckmann, Martin ; Joublin, Frank ; Goerick, Christian (2008)
Hierarchical Spectro-Temporal Features for Robust Speech Recognition.
2008 IEEE International Conference on Acoustics, Speech, and Signal Processing. Las Vegas, USA (30.03.-04.04.2008)
doi: 10.1109/ICASSP.2008.4518635
Conference or Workshop Item, Bibliographie

Abstract

Previously we presented an auditory-inspired feed-forward architecture which achieves good performance in noisy conditions on a segmented word recognition task. In this paper we propose to use a modified version of this hierarchical model to generate features for standard hidden Markov models. To obtain these features we firstly compute the spectrograms using a Gammatone filterbank. A filtering over the channels permits to enhance the formant frequencies which are afterwards detected using Gabor-like receptive fields. Then the responses of the receptive fields are combined to complex features which span the whole frequency range and extend over three different time windows. The features have been evaluated on a single digit recognition task. The results show that their combination with MFCCs or RASTA features yields improved recognition scores in noise

Item Type: Conference or Workshop Item
Erschienen: 2008
Creators: Domont, Xavier ; Heckmann, Martin ; Joublin, Frank ; Goerick, Christian
Type of entry: Bibliographie
Title: Hierarchical Spectro-Temporal Features for Robust Speech Recognition
Language: English
Date: 12 May 2008
Publisher: IEEE
Book Title: 2008 IEEE International Conference on Acoustics, Speech, and Signal Processing: Proceedings
Event Title: 2008 IEEE International Conference on Acoustics, Speech, and Signal Processing
Event Location: Las Vegas, USA
Event Dates: 30.03.-04.04.2008
DOI: 10.1109/ICASSP.2008.4518635
Abstract:

Previously we presented an auditory-inspired feed-forward architecture which achieves good performance in noisy conditions on a segmented word recognition task. In this paper we propose to use a modified version of this hierarchical model to generate features for standard hidden Markov models. To obtain these features we firstly compute the spectrograms using a Gammatone filterbank. A filtering over the channels permits to enhance the formant frequencies which are afterwards detected using Gabor-like receptive fields. Then the responses of the receptive fields are combined to complex features which span the whole frequency range and extend over three different time windows. The features have been evaluated on a single digit recognition task. The results show that their combination with MFCCs or RASTA features yields improved recognition scores in noise

Additional Information:

Print-ISBN: 978-1-4244-1483-3

Divisions: 18 Department of Electrical Engineering and Information Technology
18 Department of Electrical Engineering and Information Technology > Institut für Automatisierungstechnik und Mechatronik
18 Department of Electrical Engineering and Information Technology > Institut für Automatisierungstechnik und Mechatronik > Control Methods and Robotics (from 01.08.2022 renamed Control Methods and Intelligent Systems)
Date Deposited: 16 Aug 2010 14:32
Last Modified: 02 May 2023 11:41
PPN:
Export:
Suche nach Titel in: TUfind oder in Google
Send an inquiry Send an inquiry

Options (only for editors)
Show editorial Details Show editorial Details