Presentation | 2018-03-14 Silent Japanese Single Syllable Recognition using Similarities of Muscular Activity Transitions between EMG Channels Hidetoshi Nagai, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | In inaudible speech recognition based on surface EMG, recognition of consonants is one of the important tasks. When uttering some kinds of consonants, muscle activities change rapidly in a short time. So, in order to capture the transitions using a sequence of windows, it is necessary to narrow the windows and to reduce the shift size of windows. In that case, the extracted feature values becomes vulnerable to signal changes like noise and fluctuation of utterance speed. Also, not all muscles contribute equivalently to various utterance. The feature values extracted from the EMGs of muscles that are not contributing is the same as the noise, and will adversely affect the recognition result. However, it is difficult to judge the degree of contribution of the channel from the signal of the channel. At utterance, muscles work cooperatively according to the utterance. Therefore, there is a high possibility that features that are effective for recognition may exist in the state of cooperation of the muscles. In this paper, we propose a feature based on the similarity of the muscle activity transitions of the channel pair and the importance of the pair. And using the feature as the recognition parameter, we tried to recognize an isolated Japanese syllable which is a single vowel or a sequence of a consonant and a vowel. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | inaudible speech recognition / silent speech recognition / surface EMG / wavelet analysis / wavelets' center-of-balance / center-of-balance transition method |
Paper # | MBE2017-104 |
Date of Issue | 2018-03-06 (MBE) |
Conference Information | |
Committee | MBE / NC |
---|---|
Conference Date | 2018/3/13(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Kikai-Shinko-Kaikan Bldg. |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | ME, general |
Chair | Kazuki Nakajima(Univ. of Toyama) / Masafumi Hagiwara(Keio Univ.) |
Vice Chair | Masaki Kyoso(TCU) / Yutaka Hirata(Chubu Univ.) |
Secretary | Masaki Kyoso(Toyama Pref. Univ.) / Yutaka Hirata(Kindai Univ.) |
Assistant | Kim Juhyon(Univ. of Toyama) / Takumi Kobayashi(YNU) / Yoshihisa Shinozawa(Keio Univ.) / Keiichiro Inagaki(Chubu Univ.) |
Paper Information | |
Registration To | Technical Committee on ME and Bio Cybernetics / Technical Committee on Neurocomputing |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Silent Japanese Single Syllable Recognition using Similarities of Muscular Activity Transitions between EMG Channels |
Sub Title (in English) | |
Keyword(1) | inaudible speech recognition |
Keyword(2) | silent speech recognition |
Keyword(3) | surface EMG |
Keyword(4) | wavelet analysis |
Keyword(5) | wavelets' center-of-balance |
Keyword(6) | center-of-balance transition method |
1st Author's Name | Hidetoshi Nagai |
1st Author's Affiliation | Kyushu Institute of Technology(Kyutech) |
Date | 2018-03-14 |
Paper # | MBE2017-104 |
Volume (vol) | vol.117 |
Number (no) | MBE-507 |
Page | pp.pp.119-124(MBE), |
#Pages | 6 |
Date of Issue | 2018-03-06 (MBE) |