In-situ adaptive speech enhancement using directional microphone applied to magnetic resonance imaging

Guohua Sun, Mingfeng Li, Brent W. Rudd, Teik C. Lim, Jeffrey Osterhage, Elizabeth M. Fugate, Jing Huei Lee

Research output: Contribution to conferencePaperpeer-review

Abstract

Magnetic resonance imaging (MRI) is increasingly applied for medical diagnosis and biomedical research; however, the acoustic noise during the MRI operation remains one of the major concerns. In particular, speech communication through intercoms from patients to MRI operators is severely contaminated by the annoying noise, and requires treatment for improving the intelligibility. This study proposes an enhanced adaptive speech enhancement (ASE) system, utilizing the idea of a differential directional microphone for measuring and distinguishing the reference and desired signals, for the MRI environment to improve the speech communication. Hence, the adaptive filter is expected to remove the periodic MRI acoustic noises with limited distortion on the recovered speech signals. In-situ experiment was conducted to evaluate the system performance. Results show that the speech signal is noticeably recovered after treated with the ASE system. Furthermore, a non-adaptive spectra subtraction method is also adopted to process the abovementioned results such that additional reductions were achieved for the non-coherent MRI noises.

Original languageEnglish (US)
StatePublished - 2015
Externally publishedYes
Event44th International Congress and Exposition on Noise Control Engineering, INTER-NOISE 2015 - San Francisco, United States
Duration: Aug 9 2015Aug 12 2015

Conference

Conference44th International Congress and Exposition on Noise Control Engineering, INTER-NOISE 2015
Country/TerritoryUnited States
CitySan Francisco
Period8/9/158/12/15

All Science Journal Classification (ASJC) codes

  • Acoustics and Ultrasonics

Fingerprint

Dive into the research topics of 'In-situ adaptive speech enhancement using directional microphone applied to magnetic resonance imaging'. Together they form a unique fingerprint.

Cite this