The Impact of Speech Anonymization on Pathology and Its Limits

Overview

• This is the official repository of the paper The Impact of Speech Anonymization on Pathology and Its Limits.

Abstract

....

Prerequisites

The software is developed in Python 3.9. For the deep learning, the PyTorch 1.13 framework is used.

Main Python modules required for the software can be installed from ./requirements:

$ conda env create -f requirements.yaml
$ conda activate pathology_anonym

Note: This might take a few minutes.

Code structure

Our source code for training and evaluation of the deep neural networks, speech analysis and preprocessing are available here.

1. Everything can be run from ./PathologyAnonym_main.py.

• The data preprocessing parameters, directories, hyper-parameters, and model parameters can be modified from ./configs/config.yaml.
• Also, you should first choose an experiment name (if you are starting a new experiment) for training, in which all the evaluation and loss value statistics, tensorboard events, and model & checkpoints will be stored. Furthermore, a config.yaml file will be created for each experiment storing all the information needed.
• For testing, just load the experiment which its model you need.

2. The rest of the files:

• ./data/ directory contains all the data preprocessing, and loading files.
• ./mcAdams_Anonym/ directory contains all the files for anonymization using McAdams coefficient method.
• ./Pitch_Anonym/ directory contains all the files for anonymization using training-based randomized pitch shift + HiFi-GAN method.
• ./PathologyAnonym_Train_Valid.py contains the training and validation processes.
• ./pathanonym_Prediction.py all the prediction and testing processes.
• For EER calculation you should use either of the anonymization methods' folders based on your need.

---

In case you use this repository, please cite the original paper:

Tayebi Arasteh S, Arias-Vergara T, Perez-Toro PA, et al. The Impact of Speech Anonymization on Pathology and Its Limits. arXiv:2404.08064 (2024).

BibTex

@article {pathology_anonym,
  author = {Tayebi Arasteh, Soroosh and Arias-Vergara, Tomas and Perez-Toro, Paula Andrea and Weise, Tobias and Packhäuser, Kai and Schuster, Maria and Noeth, Elmar and Maier, Andreas and Yang, Seung Hee},
  title = {The Impact of Speech Anonymization on Pathology and Its Limits},
  year = {2024},
  journal = {arXiv:2404.08064},
  doi = {10.48550/arXiv.2404.08064}
}