Conference
Automatic Emotions Recognition in Audio Signals
| العنوان: | Automatic Emotions Recognition in Audio Signals |
|---|---|
| المؤلفون: | Tavernier, Emilie, Marechal, Catherine, Conteville, Laurie, Bougueroua, Lamine |
| المساهمون: | École d'ingénieur généraliste en informatique et technologies du numérique (EFREI), AllianSTIC-EFREI, Université Toulouse III - Paul Sabatier UPS |
| المصدر: | JETSAN 2021 - Colloque en Télésanté et dispositifs biomédicaux - 8ème édition ; https://hal.archives-ouvertes.fr/hal-03501193 ; JETSAN 2021 - Colloque en Télésanté et dispositifs biomédicaux - 8ème édition, Université Toulouse III - Paul Sabatier [UPS], May 2021, Toulouse, Blagnac, France |
| بيانات النشر: | HAL CCSD |
| سنة النشر: | 2021 |
| المجموعة: | Archive ouverte HAL (Hyper Article en Ligne, CCSD - Centre pour la Communication Scientifique Directe) |
| مصطلحات موضوعية: | Speech Emotion Recognition, IA, SVM, MFCC, [SPI.SIGNAL]Engineering Sciences [physics]/Signal and Image processing, [SPI.TRON]Engineering Sciences [physics]/Electronics |
| جغرافية الموضوع: | Toulouse, Blagnac, France |
| الوصف: | International audience ; Automatic emotions recognition is a great research subjectnowadays. It applies in different fields as robotic, data analysis, E-health. We can use different supports for emotions recognition: text,vision and voice. These three sources complement each other toanalyse the emotional state of someone. In the field of healthespecially, this can be extremely helpful to detect pathologies early,like depression, so that doctors can prescribe rapidly appropriatehealthcare.This paper deals with classification methods of eight emotions withvoice recordings. Machine learning algorithms are used.Particularly, the quadratic Support Vector Machines (SVM) gives thebest results. We use supervised training (voice recordings are labelledaccording to the emotion) and features as the maximum, the means,and the variance of various acoustic properties. We use labelleddatabases of English voice recordings for our emotionalclassification study. The results depend on the number of emotions tobe detected and on the type of emotion. |
| نوع الوثيقة: | conference object |
| اللغة: | English |
| Relation: | hal-03501193; https://hal.archives-ouvertes.fr/hal-03501193; https://hal.archives-ouvertes.fr/hal-03501193/document; https://hal.archives-ouvertes.fr/hal-03501193/file/Automatic_Emotions_Recognition_in_audio_signals.pdf |
| الاتاحة: | https://hal.archives-ouvertes.fr/hal-03501193 https://hal.archives-ouvertes.fr/hal-03501193/document https://hal.archives-ouvertes.fr/hal-03501193/file/Automatic_Emotions_Recognition_in_audio_signals.pdf |
| Rights: | info:eu-repo/semantics/OpenAccess |
| رقم الانضمام: | edsbas.D5333AB6 |
| قاعدة البيانات: | BASE |
| الوصف غير متاح. |