| 1 |
Tone realization in Mandarin speech: a large corpus based study of disyllabic words
|
|
|
|
In: The 12th International Symposium on Chinese Spoken Language Processing (ISCSLP 2021) ; https://hal.archives-ouvertes.fr/hal-03153413 ; The 12th International Symposium on Chinese Spoken Language Processing (ISCSLP 2021), Jan 2021, Hong Kong, China (2021)
|
|
BASE
|
|
Show details
|
|
|
| 2 |
Modelling the realization of variable word-final schwa in Standard French
|
|
|
|
In: 43. Jahrestagung der Deutschen Gesellschaft für Sprachwissenschaft (DGfS) ; https://hal.sorbonne-universite.fr/hal-03439307 ; 43. Jahrestagung der Deutschen Gesellschaft für Sprachwissenschaft (DGfS), Feb 2021, Freiburg, Germany (2021)
|
|
BASE
|
|
Show details
|
|
|
| 3 |
Modeling the effect of military oxygen masks on speech characteristics
|
|
|
|
In: Interspeech 2021 ; https://hal.archives-ouvertes.fr/hal-03325087 ; Interspeech 2021, Aug 2021, Brno, Czech Republic (2021)
|
|
BASE
|
|
Show details
|
|
|
| 4 |
Word-Initial Voicing Alternations in Five Romance Languages
|
|
|
|
In: Phonetics and Phonology in Europe - PaPE ; https://hal.sorbonne-universite.fr/hal-03439362 ; Phonetics and Phonology in Europe - PaPE, Jun 2021, Barcelona, Spain (2021)
|
|
BASE
|
|
Show details
|
|
|
| 5 |
Synchronic Fortition in Five Romance Languages? A Large Corpus-Based Study of Word-Initial Devoicing
|
|
|
|
In: Proceedings of Interspeech ; Interspeech 2021 ; https://hal.sorbonne-universite.fr/hal-03339852 ; Interspeech 2021, Aug 2021, Brno, Czech Republic. pp.996-1000, ⟨10.21437/Interspeech.2021-939⟩ (2021)
|
|
BASE
|
|
Show details
|
|
|
| 6 |
A corpus-based study of the distribution of word-final schwa in Standard French and what it teaches us about its phonological status
|
|
|
|
In: ISSN: 2385-4138 ; Isogloss. Open Journal of Romance Linguistics ; https://hal.sorbonne-universite.fr/hal-03499017 ; Isogloss. Open Journal of Romance Linguistics, 2021 (2021)
|
|
BASE
|
|
Show details
|
|
|
| 7 |
Context, position in word and duration as predictors of voicing alternation of stops: a large-scale corpus-based study in 5 Romance Languages
|
|
|
|
In: PaPE 2021 - 4th phonetics and phonology in Europe ; https://hal.archives-ouvertes.fr/hal-03438673 ; PaPE 2021 - 4th phonetics and phonology in Europe, Jun 2021, Barcelone, Spain (2021)
|
|
BASE
|
|
Show details
|
|
|
| 8 |
Comparaison dialectométriques de parlers du Croissant avec d'autres parlers d'oc et d'oïl
|
|
|
|
In: Le Croissant linguistique entre oc, oïl et francoprovençal : des mots à la grammaire, des parlers aux aires ; https://hal.archives-ouvertes.fr/hal-03318765 ; Le Croissant linguistique entre oc, oïl et francoprovençal : des mots à la grammaire, des parlers aux aires, 2021 (2021)
|
|
BASE
|
|
Show details
|
|
|
| 9 |
End-to-End Speech Emotion Recognition: Challenges of Real-Life Emergency Call Centers Data Recordings
|
|
|
|
In: ISBN: 978-1-6654-0019-0 ; 2021 9th International Conference on Affective Computing and Intelligent Interaction (ACII) ; https://hal.archives-ouvertes.fr/hal-03405970 ; 2021 9th International Conference on Affective Computing and Intelligent Interaction (ACII), Sep 2021, Nara, Japan ; https://www.acii-conf.net/2021/ (2021)
|
|
Abstract:
International audience ; Recognizing a speaker's emotion from their speech can be a key element in emergency call centers. End-to-end deep learning systems for speech emotion recognition now achieve equivalent or even better results than conventional machine learning approaches. In this paper, in order to validate the performance of our neural network architecture for emotion recognition from speech, we first trained and tested it on the widely used corpus accessible by the community, IEMOCAP. We then used the same architecture as the real life corpus, CEMO, composed of 440 dialogs (2h16m) from 485 speakers. The most frequent emotions expressed by callers in these real life emergency dialogues are fear, anger and positive emotions such as relief. In the IEMOCAP general topic conversations, the most frequent emotions are sadness, anger and happiness. Using the same end-to-end deep learning architecture, an Unweighted Accuracy Recall (UA) of 63% is obtained on IEMOCAP and a UA of 45.6% on CEMO, each with 4 classes. Using only 2 classes (Anger, Neutral), the results for CEMO are 76.9% UA compared to 81.1% UA for IEMOCAP. We expect that these encouraging results with CEMO can be improved by combining the audio channel with the linguistic channel. Real-life emotions are clearly more complex than acted ones, mainly due to the large diversity of emotional expressions of speakers. Index Terms-emotion detection, end-to-end deep learning architecture, call center, real-life database, complex emotions.
|
|
Keyword:
[INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]; [INFO.INFO-CL]Computer Science [cs]/Computation and Language [cs.CL]; [INFO.INFO-SD]Computer Science [cs]/Sound [cs.SD]; [STAT.ML]Statistics [stat]/Machine Learning [stat.ML]; deep learning system; emergency call center; real life; recurrent neural network; speech emotion recognition; temporal neural networks
|
|
URL: https://hal.archives-ouvertes.fr/hal-03405970/document
https://hal.archives-ouvertes.fr/hal-03405970/file/main.pdf
https://hal.archives-ouvertes.fr/hal-03405970
|
|
BASE
|
|
Hide details
|
|
|
| 10 |
Fine phonetic details for DM disambiguation in French: a corpus-based investigation ...
|
|
|
|
BASE
|
|
Show details
|
|
|
| 11 |
A corpus-based study of the distribution of word-final schwa in Standard French and what it teaches us about its phonological status
|
|
|
|
BASE
|
|
Show details
|
|
|
|
|
