报告题目: |
Man-Machine Interface: Dynamic pitch perception with cochlear implants |
报告人: |
Xin Luo |
|
Ph.D.
Assistant Professor of the Department of Speech, Language,
and Hearing Sciences at Purdue University
|
报告时间: |
2011-06-14 15:00 |
报告地点: |
医学院B321 |
主办单位: |
医学院 |
简介: |
Cochlear implants (CIs) partially restore hearing sensation to profoundly deaf people via direct electric stimulation of the surviving auditory neurons. Although many CI users understand speech well in quiet, most of them have great difficulty in music melody recognition and supra-segmental speech tasks (e.g., recognition of lexical tones, speech prosody, and talker identity), due to poor access to dynamic pitch cues. In this talk, different strategies to enhance CI users’ pitch contour perception will be discussed. The first approach represents pitch contours in the loudness domain to compensate for poor pitch coding. A pre-processing strategy modifying speech amplitude envelope to better resemble its pitch contour has been shown to significantly improve tone recognition with CI. The second approach provides more salient pitch cues via residual low-frequency acoustic hearing. Combined electric and acoustic stimulation has been found to benefit tone recognition, talker identification, and vocal emotion recognition, among others. Finally, the third approach uses both place and temporal coding to better transmit pitch contours to CI users. Our results showed that pitch contours can be encoded using either continuous current steering or time-varying modulation frequencies, and pitch contour identification can be enhanced with consistent place and temporal cues. |
|
