Project

The main objective of this project is to predict emotion from audio

Software specifications:

These projects where designed and tested on Lemans-university servers which are equiped with GPU graphic cards.

Train/Test Data

two German emotional corpuses

1. EMO-DB
2. AIBO

Classes:

1. Anger (angry, touchy, et reprimanding)
2. Emphatic
3. Neutral
4. Positive (motherese et joyful)
5. Rest
6. Negative (angry, touchy, reprimanding, et emphatic)
7. IDLe (all non-negative states)

Author

Hichem MAIZA

Acknoldgements:

• Thank you opensmile for the great mfcc extraction script
• Thank you PD Dr.-Ing. habil. Stefan Steidl for the great aibo corpus
• Thank you Felix Burkhardt, Miriam Kienast, Astrid Paeschke and Benjamin Weiss for the great Emo-DB corpus