| Presentation | 2015-10-16 Switch-To-Speech Communication Aid System Using WFST and Low Latency Search Algorithm Fuming Fang, Takahiro Shinozaki, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | To establish a communication method for the patients who have lost nearly all of their voluntary movementincluding speech, we have proposed a switch-to-speech communication aid system. Since this system candirectly convert the motion sequence into string of characters using an input protocol, it is portable compared tothe other system that uses a large display. In order to produce efficient and easily-learned protocol, we have alsoproposed an input protocol recommendation method. This method produces a protocol that has less combination ofmotions to be remembered by mapping multiple Kana characters to the same motion sequence. To compensate forthe introduced ambiguity of converting input motion sequence to Kana sequence, a word N-gram language modelis used. However, the method has a problem that there is a large delay before intended Kana is output after thecorresponding motion sequence is input since it requires to search a long language context. In this paper, we firstpropose to adopt a Kana N-gram instead of a high precision word N-gram. Secondly, we propose a low latency searchalgorithm at the expense of accuracy, which compulsorily outputs the best path if the delay exceeds a pre-definedvalue. These two methods significantly reduce the delay from 4.6 Kana to 0.9 Kana, while the accuracy is slightlyreduced from 97.2% to 92.8%. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Communication aid system / WFST / low latency search / N-gram |
| Paper # | SP2015-68 |
| Date of Issue | 2015-10-08 (SP) |
| Conference Information | |
| Committee | SP |
|---|---|
| Conference Date | 2015/10/15(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Kobe Univ. |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Speech interface, Synthesis, Dialogue, Application system, etc. |
| Chair | Kazunori Mano(Shibaura Inst. of Tech.) |
| Vice Chair | Norihide Kitaoka(Tokushima Univ.) |
| Secretary | Norihide Kitaoka(Tokyo City Univ.) |
| Assistant | Takashi Nose(Tohoku Univ.) / Taichi Asami(NTT) |
| Paper Information | |
| Registration To | Technical Committee on Speech |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Switch-To-Speech Communication Aid System Using WFST and Low Latency Search Algorithm |
| Sub Title (in English) | |
| Keyword(1) | Communication aid system |
| Keyword(2) | WFST |
| Keyword(3) | low latency search |
| Keyword(4) | N-gram |
| 1st Author's Name | Fuming Fang |
| 1st Author's Affiliation | Tokyo Institute of Technology(Tokyo Tech) |
| 2nd Author's Name | Takahiro Shinozaki |
| 2nd Author's Affiliation | Tokyo Institute of Technology(Tokyo Tech) |
| Date | 2015-10-16 |
| Paper # | SP2015-68 |
| Volume (vol) | vol.115 |
| Number (no) | SP-253 |
| Page | pp.pp.51-56(SP), |
| #Pages | 6 |
| Date of Issue | 2015-10-08 (SP) |