Speech recognition in noisy environments: a survey
Speech Communication
Robust automatic speech recognition with missing and unreliable acoustic data
Speech Communication
Outlier Detection Using Classifier Instability
SSPR '98/SPR '98 Proceedings of the Joint IAPR International Workshops on Advances in Pattern Recognition
Novelty detection: a review—part 1: statistical approaches
Signal Processing
Computational Auditory Scene Analysis: Principles, Algorithms, and Applications
Computational Auditory Scene Analysis: Principles, Algorithms, and Applications
Integrating computational auditory scene analysis and automatic speech recognition
Integrating computational auditory scene analysis and automatic speech recognition
The application of hidden Markov models in speech recognition
Foundations and Trends in Signal Processing
A Bayesian estimation approach for speech enhancement using hiddenMarkov models
IEEE Transactions on Signal Processing
Separation of speech from interfering sounds based on oscillatory correlation
IEEE Transactions on Neural Networks
Monaural speech segregation based on pitch tracking and amplitude modulation
IEEE Transactions on Neural Networks
Proceedings of the Second Symposium on Information and Communication Technology
Robust speech recognition based on binaural speech enhancement system as a preprocessing step
Proceedings of the Third Symposium on Information and Communication Technology
| Hi-index | 0.00 |
Missing-data methods attempt to improve robust speech recognition by distinguishing between reliable and unreliable data in the time-frequency (T-F) domain. Such methods require a binary mask to label speech-dominant T-F regions of a noisy speech signal as reliable and the rest as unreliable. Current methods for computing the mask are based mainly on bottom-up cues such as harmonicity and produce labeling errors that degrade recognition performance. In this paper, we propose a two-stage recognition system that combines bottom-up and top-down cues in order to simultaneously improve both mask estimation and recognition accuracy. First, an n-best lattice consistent with a speech separation mask is generated. The lattice is then re-scored by expanding the mask using a model-based hypothesis test to determine the reliability of individual T-F units. Systematic evaluations of the proposed system show significant improvement in recognition performance compared to that using speech separation alone.