Journal article

BARTOS Anthony L., CIPR Tomáš, NELSON Douglas J., SCHWARZ Petr, BANOWETZ John and JERABEK Ladislav. Noise-robust speech triage. The Journal of the Acoustical Society of America. 2018, vol. 143, no. 4, pp. 2313-2320. ISSN 1520-8524. Available from: https://asa.scitation.org/doi/10.1121/1.5031029
Publication language:english
Original title:Noise-robust speech triage
Title (cs):Třídění řeči odolné vůči šumu
Pages:2313-2320
Place:US
Year:2018
URL:https://asa.scitation.org/doi/10.1121/1.5031029
Journal:The Journal of the Acoustical Society of America, Vol. 143, No. 4, US
ISSN:1520-8524
DOI:10.1121/1.5031029
URL:http://www.fit.vutbr.cz/research/groups/speech/publi/2018/A_L_Bartos_JASMAN_vol_143iss_42313_1.pdf [PDF]
Keywords
speech algorithms, noisy environments, multiple speaker identification
Annotation
A method is presented in which conventional speech algorithms are applied, with no modifications, to improve their performance in extremely noisy environments. It has been demonstrated that, for eigen-channel algorithms, pre-training multiple speaker identification (SID) models at a lattice of signal-to-noise-ratio (SNR) levels and then performing SID using the appropriate SNR dependent model was successful in mitigating noise at all SNR levels. In those tests, it was found that SID performance was optimized when the SNR of the testing and training data were close or identical. In this current effort multiple i-vector algorithms were used, greatly improving both processing throughput and equal error rate classification accuracy. Using identical approaches in the same noisy environment, performance of SID, language identification, gender identification, and diarization were significantly improved. A critical factor in this improvement is speech activity detection (SAD) that performs reliably in extremely noisy environments, where the speech itself is barely audible. To optimize SAD operation at all SNR levels, two algorithms were employed. The first maximized detection probability at low levels (10 dB SNR < 10 dB) using just the voiced speech envelope, and the second exploited features extracted from the original speech to improve overall accuracy at higher quality levels (SNR10 dB).
BibTeX:
@ARTICLE{
   author = {L. Anthony Bartos and Tom{\'{a}}{\v{s}} Cipr and J. Douglas
	Nelson and Petr Schwarz and John Banowetz and Ladislav
	Jerabek},
   title = {Noise-robust speech triage},
   pages = {2313--2320},
   journal = {The Journal of the Acoustical Society of America},
   volume = {143},
   number = {4},
   year = {2018},
   ISSN = {1520-8524},
   doi = {10.1121/1.5031029},
   language = {english},
   url = {http://www.fit.vutbr.cz/research/view_pub.php?id=11716}
}

Your IPv4 address: 54.224.60.122
Switch to IPv6 connection

DNSSEC [dnssec]