CITATION — REFERENCE ENTRY
An algorithm to improve speech recognition in noise for hearing-impaired listeners — The Journal of the Acoustical Society of America
- Key
- healy2013-hearing-impaired
- Authors
- Healy, Eric W.; Yoho, Sarah E.; Wang, Yuxuan; Wang, DeLiang
- Issued
- 2013-10
- Type
- article-journal
- Container
- The Journal of the Acoustical Society of America
- Volume
- 134
- Issue
- 4
- Pages
- 3029-3038
Raw CSL JSON
{
"DOI": "10.1121/1.4820893",
"URL": "https://pnlwang.github.io/papers/HYWW.jasa13.pdf",
"PMID": "24116438",
"page": "3029-3038",
"type": "article-journal",
"PMCID": "PMC3799726",
"issue": "4",
"title": "An algorithm to improve speech recognition in noise for hearing-impaired listeners",
"author": [
{
"given": "Eric W.",
"family": "Healy"
},
{
"given": "Sarah E.",
"family": "Yoho"
},
{
"given": "Yuxuan",
"family": "Wang"
},
{
"given": "DeLiang",
"family": "Wang"
}
],
"issued": {
"date-parts": [
[
2013,
10
]
]
},
"volume": "134",
"container-title": "The Journal of the Acoustical Society of America"
}
Claims
-
Deep neural networks were used as classifiers because they had been shown to outperform earlier classifiers.
"Wang and Wang (2013) showed that deep neural networks (DNNs) outperform both of these. DNNs were therefore employed as classifiers in the current study."
-
Unlike the ideal binary mask, which requires prior knowledge of the premixed signals, the masks in this study were estimated by an algorithm trained on speech not used in testing.
"Unlike the ideal binary mask, which requires prior knowledge of the premixed signals, the masks used to segregate speech from noise in the current study were estimated by training the algorithm on speech not used during testing."
-
The authors describe it as the first monaural algorithm with demonstrated speech intelligibility improvements for hearing-impaired listeners in background noise.
"To our knowledge, this is the first monaural algorithm that provides demonstrated speech intelligibility improvements for HI listeners in background noise."
-
Intelligibility increased after processing in all conditions, more for hearing-impaired listeners, for the modulated background and for the least favorable SNRs.
"Testing using normal-hearing and HI listeners indicated that intelligibility increased following processing in all conditions. These increases were larger for HI listeners, for the modulated background, and for the least-favorable SNRs."
-
Implementation in hearing devices is a long-term goal, and the algorithm was far from ready to implement.
"We stress that this goal is long term and that the current algorithm is far from ready to implement."
Available in