Discriminative feature domains for reverberant acoustic environments
File(s)
Accepted version
Author(s)
Papayiannis, C
Evers, C
Naylor, PA
Type
Conference Paper
Abstract
Several speech processing and audio data-mining applications
rely on a description of the acoustic environment as a feature
vector for classification. The discriminative properties of the
feature domain play a crucial role in the effectiveness of these
methods. In this work, we consider three environment iden-
tification tasks and the task of acoustic model selection for
speech recognition. A set of acoustic parameters and Ma-
chine Learning algorithms for feature selection are used and
an analysis is performed on the resulting feature domains for
each task. In our experiments, a classification accuracy of
100% is achieved for the majority of tasks and the Word Er-
ror Rate is reduced by 20.73 percentage points for Automatic
Speech Recognition when using the resulting domains. Ex-
perimental results indicate a significant dissimilarity in the
parameter choices for the composition of the domains, which
highlights the importance of the feature selection process for
individual applications.
rely on a description of the acoustic environment as a feature
vector for classification. The discriminative properties of the
feature domain play a crucial role in the effectiveness of these
methods. In this work, we consider three environment iden-
tification tasks and the task of acoustic model selection for
speech recognition. A set of acoustic parameters and Ma-
chine Learning algorithms for feature selection are used and
an analysis is performed on the resulting feature domains for
each task. In our experiments, a classification accuracy of
100% is achieved for the majority of tasks and the Word Er-
ror Rate is reduced by 20.73 percentage points for Automatic
Speech Recognition when using the resulting domains. Ex-
perimental results indicate a significant dissimilarity in the
parameter choices for the composition of the domains, which
highlights the importance of the feature selection process for
individual applications.
Date Issued
2017-06-19
Date Acceptance
2016-12-18
Citation
Proc. IEEE Intl. Conf. on Acoustics, Speech and Signal Processing (ICASSP), 2017
ISSN
2379-190X
Publisher
IEEE
Journal / Book Title
Proc. IEEE Intl. Conf. on Acoustics, Speech and Signal Processing (ICASSP)
Copyright Statement
© 2017 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
Sponsor
Commission of the European Communities
Grant Number
609465
Source
2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Subjects
Science & Technology
Technology
Acoustics
Engineering, Electrical & Electronic
Engineering
Feature Selection
Machine Learning
Environment Identification
Reverberant speech recognition
EQUALIZATION
Publication Status
Published
Start Date
2017-03-05
Finish Date
2017-03-09
Coverage Spatial
New Orleans, LA