An Speech Emotion Recognition project for the Advanced Artificial Intelligence course in National University of Engineering.
We preprocessed each audio, trimming the silence, and then we extract its spectrogram as an image. We feed this image to a Convolutional Neural Networks (CNN) to predict the emotion.
We used 2 datasets to compare the results : RAVDESS (https://zenodo.org/record/1188976) and a dataset created by us.
- Hans Martin Acha Carranza
- Diego Hurtado de Mendoza González Zúñiga
- Jair Puican Cuadros