Speech Emotion Recognition models implementation with TensorFlow and Keras
On this repository I share my studies and a simple implementation of speech emotion recognition models that reached a very high-competitive accuracy comparing with the state-of-the-art papers available nowadays.
I used IEMOCAP dataset for the Multimodal application and RAVDESS dataset for the Audio only model application.
-
RAVDESS Dataset can be found at: https://zenodo.org/record/1188976#.XiQ7NcgzaUk
-
IEMOCAP Dataset can be found at: https://sail.usc.edu/iemocap/ (license agreement is required)
On both models, MFCC (n = 40) was used as audio input feature.