/Pretrained_audio_neural_networks_emotion_recognition

Code for the paper 'Pretrained audio neural networks for Speech emotion recognition in Portuguese'

Primary LanguageJupyter NotebookGNU General Public License v3.0GPL-3.0

Pretrained_audio_neural_networks_emotion_recognition

This repository contains code for the paper 'Pretrained audio neural networks for Speech emotion recognition in Portuguese' published in the Workshop Automatic Speech Recognition for spontaneous and prepared speech & Speech Emotion Recognition in Portuguese in 2022.

The folder MFCC_gram_Transformers contains the code used for the Transformer Encoder units in the paper. It is based on the code found at http://nlp.seas.harvard.edu/2018/04/03/attention.html; Further information is found within the folder

The folder Pretrained_audio_neural_networks contains the code used for the PANNs in the paper. It uses the repository: https://github.com/qiuqiangkong/audioset_tagging_cnn which is the original repository for the PANNs and adds a few utility functions to run PANNs for speech emotion recognition. Further information is found within the folder

For questions regarding the implementation please contact: marcelomatheusgauy@gmail.com