Robust automatic speech recognition with missing and unreliable acoustic data

By Martin P. Cooke, Phil D. Green, Ljubomir Josifovski, Ascension Vizinho. Speech Communication, 2001

746 citations recorded by the source corpus.

Subfields: Computational Mechanics, Signal Processing

View Paper: https://www.semanticscholar.org/paper/15750cad2db8b6419fa252a712c3aa9eeda564d3

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

1. Deliang Wang
2. Jinyu Li
3. John F. Woodruff
4. Nicoleta Roman
5. Yifan Gong
6. Shahla Parveen
7. Daniel Pullella
8. Roberto B. Togneri

In Signal Processing

1. Jinyu Li
2. Deliang Wang
3. Yifan Gong
4. Dong Yu
5. Li Deng
6. Arun Narayanan
7. Nicoleta Roman
8. John F. Woodruff

Citations in the corpus

• Continuous latent variable models for dimensionality reduction and sequential data reconstruction (2001, background): “8, reconstructing the speech and the classifying it as usual is worse(35) than attempting directly to classify conditionally on the present variables, as Cooke et al. (2001) have demonstrated empirically.”
• Speech segregation based on sound localization (2001, background): “Recent investigations also show that robust results can be obtained using similar binary masks as front-end processors to automatic speech recognizers [12].”
• An Overview of Speaker Identification: Accuracy and Robustness Issues (2011, methodology): “For use as a pure missing feature estimator, previous work on missing data speech recognition reports that the 0 dB criterion is superior to the negative energy criterion [74], [87].”
• Speech enhancement with missing data techniques using recurrent neural networks (2004, methodology): “The operation of automatic speech recognition (ASR) systems, telecommunication applications and cellular phones under varying background noise conditions has made speech enhancement an interesting area of research for the past two decades [1-8].”
• Monaural speech segregation based on pitch tracking and amplitude modulation (2004, neutral): “An ideal mask is also similar to an a priori mask used in a recent ASR study [9], which yields excellent recognition performance.”
• We Can Hear You with Wi-Fi! (2014, background): “That is, WiHear builds a wireless-based provocation dictionary for automatic speech recognition system [22].”
• Speech intelligibility in reverberation with ideal binary masking: effects of early reflections and signal-to-noise ratio threshold. (2013, background): “Ideal binary T-F masks have proven to provide highly effective front ends for robust speech recognition (Cooke et al. , 2001; Roman et al. , 2003; Srinivasan et al. , 2006; Srinivasan and Wang, 2010) as well as to improve speech intelligibility for both normal-hearing and hearing-impaired listeners…”
• An Overview of Noise-Robust Automatic Speech Recognition (2014, background): “These unreliable cells are either ignored or fi lled in by estimates of their putative values in subsequent processing [252], [253].”