| Project/Area Number |
12650375
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
情報通信工学
|
|---|
| Research Institution | Shizuoka University |
|---|
Principal Investigator |
KITAZAWA Shigeyoshi Shizuoka University, Computer Science, Professor, 情報学部, 教授 (00109018)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
KITAMURA Tatsuya Shizuoka University, Computer Science, Research Associate, 情報学部, 助手 (60293594)
|
|---|
| Project Period (FY) |
2000 – 2001
|
| Project Status |
Completed (Fiscal Year 2001)
|
| Budget Amount *help |
¥3,300,000 (Direct Cost: ¥3,300,000)
Fiscal Year 2001: ¥1,400,000 (Direct Cost: ¥1,400,000)
Fiscal Year 2000: ¥1,900,000 (Direct Cost: ¥1,900,000)
|
|---|
| Keywords | peripheral auditory model / cochlear implant / multi-pulse coding / stimulated sound / acoustic simulation / interview / 逆解法 / Meddisモデル |
|---|
| Research Abstract |
A sound (environmental sound is also included and considered) is efficiently changed into the electrode pulses of 20 channels of a cochlear implant. For this purpose, expanding an algorithm of the multi-pulse voice coding usually used by the voice coding system, it was going to calculate the optimum solution in a time-frequency space to determine the optimum pulse position (on a time-axis) and the electrode position (on a frequency-axis) along the basilar-membrane in a cochlea. As a result of looking over the problem of an optimization criterion again, good results were obtained by the method of the minimum-squared-error criterion of waveform and spectrum.
The acoustic simulation system based on the physiology experiment model of the auditory-nerve firing pattern by electrode pulse stimulus was created. This can perform an accurate acoustic simulation now from an electrode stimulus signal.
The reverse process of the physiology model of a peripheral auditory system was followed, and the a
… More
coustic signal of an input was reproduced from the firing pattern of an auditory nerve. The "reverse hearing model" was created so that input sound with no distortion might be reproduced, in consideration of various restriction conditions of the nonlinear reverse model that restores basilar-membrane vibration from quantity of a nerve transmitter.
What "sound" do cochlear implantees hear? Sure enough, the "sound simulation" considered that we wanted to obtain a certain key to the question whether it corresponds to the "sound" which is hearing a cochlear implantee, and interviewed in detail dozens of people who have implanted at the Hamamatsu medical college. As a result, the following thing was understood. Imitating in an acoustic simulation is close to "sound" of the moment of the hookup to a cochlear implant. Then, a cochlear implantee has learned quickly (less than to several minutes). It becomes audible as before after several months as far as they believe. Perhaps people relates with the "sound" in ones memory that was being listened to once.
In the acoustic simulation, although the distortion of sound and mixing of noise often happened for the sake of the technical problem on signal composition, it became clear that it is entirely an unrelated problem for a cochlear implantee. Therefore, it can be said that tone-quality evaluation of cochlear-implant sound must not be mistaken by having produced an unnecessary distortion at the time of an acoustic simulation. Less
|
