DMCA
Noise Robust ASR in Reverberated Multisource Environments Applying Convolutive (2013)
| Venue: | NMF and Long Short-Term Memory,” Computer Speech and Language, Special Issue on Speech Separation and Recognition in Multisource Environments |
| Citations: | 5 - 4 self |
BibTeX
@INPROCEEDINGS{Wöllmer13noiserobust,
author = {Martin Wöllmer and Felix Weninger and Jürgen Geiger and Björn Schuller and Gerhard Rigoll},
title = {Noise Robust ASR in Reverberated Multisource Environments Applying Convolutive},
booktitle = {NMF and Long Short-Term Memory,” Computer Speech and Language, Special Issue on Speech Separation and Recognition in Multisource Environments},
year = {2013},
pages = {780--797}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
This article proposes and evaluates various methods to integrate the concept of bidirectional Long Short-Term Memory (BLSTM) temporal context modeling into a system for automatic speech recognition (ASR) in noisy and reverberated environments. Building on recent advances in Long Short-Term Memory architectures for ASR, we design a novel front-end for contextsensitive Tandem feature extraction and show how the Connectionist Temporal Classification approach can be used as a BLSTM-based back-end, alternatively to Hidden Markov Models (HMM). We combine context-sensitive BLSTM-based feature generation and speech decoding techniques with source separation by convolutive non-negative matrix factorization. Applying our speaker adapted multi-stream HMM framework that processes MFCC features from NMFenhanced speech as well as word predictions obtained via BLSTM networks and non-negative sparse classification (NSC), we obtain an average accuracy of 91.86 % on the PASCAL CHiME Challenge task at signal-to-noise ratios ranging from-6 to 9 dB. To our knowledge, this is the best result ever reported for the CHiME Challenge task.
Keyphrases
reverberated multisource environment applying convolutive noise robust asr blstm network word prediction various method source separation multi-stream hmm framework signal-to-noise ratio pascal chime challenge task mfcc feature novel front-end chime challenge task bidirectional long short-term memory blstm-based back-end convolutive non-negative matrix factorization context-sensitive blstm-based feature generation contextsensitive tandem feature extraction hidden markov model recent advance nmfenhanced speech connectionist temporal classification approach long short-term memory average accuracy temporal context automatic speech recognition non-negative sparse classification
