Speech Emotion Recognition (SER) using the EMOVOME database

This repository contains the Python implementation of the article: "EMOVOME Database: Advancing Emotion Recognition in Speech Beyond Staged Scenarios".

Citation

If you use this code in your research, please cite our work as follows:

L. Gómez-Zaragozá, R. del Amor, M. Castro-Bleda, V. Naranjo, M. Alcañiz Raya and J. Marín-Morales, "EMOVOME Database: Advancing Emotion Recognition in Speech Beyond Staged Scenarios", 2024. https://arxiv.org/abs/2403.02167

@misc{gomez2024,
  title={EMOVOME Database: Advancing Emotion Recognition in Speech Beyond Staged Scenarios},
  author={Lucía Gómez-Zaragozá and Rocío del Amor and María José Castro-Bleda and Valery Naranjo and Mariano Alcañiz Raya and Javier Marín-Morales},
  year={2024},
  eprint={2403.02167},
  archivePrefix={arXiv},
}

Files and folders

• 1_extract_features: folder with the following sub-folders
  • 1_eGeMAPS_features: code to extract the eGeMAPS features
  • 2_speech_embeddings: code to extract the speech embeddings using different pre-trained models
  • 3_combine_eGeMAPS_embeddings: code to combine the eGeMAPS features and the speech embeddings
• 2_run_SER_models: folder with the following jupyter notebooks
  • baseline_egemaps.ipynb: code to implement the baseline models based on eGeMAPS
  • model_embeddings.ipynb: code to run the models using the embeddings and/or their combination with eGeMAPS

Installation

To ensure a smooth setup, we recommend creating a new virtual environment. Follow these steps:

1. Create a new environment with Conda:

conda create --name myenv

2. Activate the newly created environment:

conda activate myenv

3. Install all dependencies from the requirements.txt file using pip:

pip3 install -r my_requirements.txt

Usage

To run the Jupyter notebooks, ensure Jupyter is installed in your environment. You can install it using:

pip install jupyter

You can then open and run the Jupyter notebooks with the following command:

jupyter notebook my_notebook.ipynb