| Presentation | 2015-03-17 Silent Season BCI : Performance of HMM Takashi ITO, Hiromi YAMAGUCHI, Ayaka YAMAGUCHI, Toshimasa YAMAZAKI, Shin-ichi FUKUZUMI, Takahiro YAMANOI, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | The purpose of this study is to generalize our previous SSBCI (Silent Speech Brain-Computer Interface) study proposed in terms of vowel recognition, The SSBCI uses both single-trial EEGs and speech signals. In this paper, we analyzed EEGs and speech signals during the silent season (SS) task only with "ham" and "natsu". First, using a Kalman filter (KF), we described the relationship between spectrograms of the speech signals and the EEGs whose dipole solutions were localized to the Broca's area. The spectrograms were transformed into vowel and consonant sequences, and these seguence were learned by Hidden Markov Model (HMM). Next, we predicted spectrograms for the SS task from the KF and calculated log-likelihood by inputting the spectrograms to the HMM. By comparing "ham"-log-likelihood and "natsu"-log-likelihood, we discriminated between "haru" and "natsu" on SS task. As a result regarding confusion matrices, the accuracy was 83% ("haru") and 67% ("natsu"). |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | BCI / silent speech / ICA / ECDL / Kalman filter / HMM |
| Paper # | MBE2014-130,NC2014-81 |
| Date of Issue |
| Conference Information | |
| Committee | NC |
|---|---|
| Conference Date | 2015/3/9(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Neurocomputing (NC) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Silent Season BCI : Performance of HMM |
| Sub Title (in English) | |
| Keyword(1) | BCI |
| Keyword(2) | silent speech |
| Keyword(3) | ICA |
| Keyword(4) | ECDL |
| Keyword(5) | Kalman filter |
| Keyword(6) | HMM |
| 1st Author's Name | Takashi ITO |
| 1st Author's Affiliation | Kyushu Institute of Technology() |
| 2nd Author's Name | Hiromi YAMAGUCHI |
| 2nd Author's Affiliation | Kyushu Institute of Technology |
| 3rd Author's Name | Ayaka YAMAGUCHI |
| 3rd Author's Affiliation | Hitachi Systems, Ltd. |
| 4th Author's Name | Toshimasa YAMAZAKI |
| 4th Author's Affiliation | Kyushu Institute of Technology |
| 5th Author's Name | Shin-ichi FUKUZUMI |
| 5th Author's Affiliation | Knowledge Discovery Research Laboratories, NEC Corp. |
| 6th Author's Name | Takahiro YAMANOI |
| 6th Author's Affiliation | Hokkai Gakuen University |
| Date | 2015-03-17 |
| Paper # | MBE2014-130,NC2014-81 |
| Volume (vol) | vol.114 |
| Number (no) | 515 |
| Page | pp.pp.- |
| #Pages | 4 |
| Date of Issue | |