| Presentation | 2019-12-06 [Poster Presentation] Analysis and Subjective Labeling for Emotional Speech Database JTES Mai Yamanaka, Takashi Nose, Yuya Chiba, Akinori Ito, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | We have constructed JTES, a prosodic balanced emotional speech database containing 50 sentences of 4 emotions of 50 men and women, totaling 20,000 utterances. However, JTES had variations in the emotional intensity and expression, which could adversely affect emotion recognition and speech synthesis. In this paper, in order to investigate the variation of speaker's emotional expression for improving the accuracy of speech emotion recognition, the subjective label (listener label) is given by the listener and the result is analyzed. The subjective labels matched the emotional label of the speaker with an accuracy of 70% but some voices that could not be correctly classified were distinguished into neutral. There is the possibility that these speeches affect emotion recognition. Crowdsourcing was used to assign large-scale subjective labels. In order to request many people to works, we introduced some speeches to easily check the reliability.However it was not possible to use it as an index for reliability. About 70% of the match rates between speaker labels and subjective labels for each labeler were 70%, but there was also one with less than 50%. The labeling bias of each labeler was vectorized and compared with the average of all labeler's vectors, and 60% of the labelers were distributed at a value of about 0.5 away from the average. Since then, the number has decreased to three. This indicates that there are individual differences in labelers. When the emotional labels assigned to each labeler were vectorized and compared with the average of all labeler vectors, the distance was found to be distributed in the form of a normal distribution from 0.2 to 0.9, and it was clear that there were large individual differences depending on the labeler. This indicates that there are individual differences in labelers. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | emotional label / subjective label / crowdsourcing |
| Paper # | SP2019-39 |
| Date of Issue | 2019-11-29 (SP) |
| Conference Information | |
| Committee | NLC / IPSJ-NL / SP / IPSJ-SLP |
|---|---|
| Conference Date | 2019/12/4(3days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | NHK Science & Technology Research Labs. |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | The 6th Natural Language Processing Symposium & The 21th Spoken Language Symposium |
| Chair | Takeshi Sakaki(Hottolink) / / Hisashi Kawai(NICT) |
| Vice Chair | Mitsuo Yoshida(Toyohashi Univ. of Tech.) / Kazutaka Shimada(Kyushu Inst. of Tech.) / / Akinobu Ri(Nagoya Inst. of Tech.) |
| Secretary | Mitsuo Yoshida(Ryukoku Univ.) / Kazutaka Shimada(NTT) / / Akinobu Ri(Kyoto Univ.) / (Waseda Univ.) |
| Assistant | Takeshi Kobayakawa(NHK) / Hiroki Sakaji(Univ. of Tokyo) / / Tomoki Koriyama(Univ. of Tokyo) / Yusuke Ijima(NTT) |
| Paper Information | |
| Registration To | Technical Committee on Natural Language Understanding and Models of Communication / Special Interest Group on Natural Language / Technical Committee on Speech / Special Interest Group on Spoken Language Processing |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | [Poster Presentation] Analysis and Subjective Labeling for Emotional Speech Database JTES |
| Sub Title (in English) | |
| Keyword(1) | emotional label |
| Keyword(2) | subjective label |
| Keyword(3) | crowdsourcing |
| 1st Author's Name | Mai Yamanaka |
| 1st Author's Affiliation | Tohoku University(Tohoku Univ.) |
| 2nd Author's Name | Takashi Nose |
| 2nd Author's Affiliation | Tohoku University(Tohoku Univ.) |
| 3rd Author's Name | Yuya Chiba |
| 3rd Author's Affiliation | Tohoku University(Tohoku Univ.) |
| 4th Author's Name | Akinori Ito |
| 4th Author's Affiliation | Tohoku University(Tohoku Univ.) |
| Date | 2019-12-06 |
| Paper # | SP2019-39 |
| Volume (vol) | vol.119 |
| Number (no) | SP-321 |
| Page | pp.pp.61-66(SP), |
| #Pages | 6 |
| Date of Issue | 2019-11-29 (SP) |