TaviG/Speaker-Identification

Speaker Identification on VoxCeleb dataset using statistical moments, covariance analysis, PCA based on covariance matrix computation and K-means

Speaker-Identification

Speaker Identification on VoxCeleb dataset using statistical moments, covariance analysis, PCA based on covariance matrix computation and K-means

Prerequisite:

sudo apt install ffmpeg

In virtual environment:

python3 -m venv

venv . venv/bin/activate

pip install requirements.txt

To run extract_ds.py:

python extract_ds.py /path/to/dataset

To run mp42wav.py

python mp42wav.py /path/to/dataset

Dataset must be downloaded and stored locally from

https://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox1.html

Audio files

All the audio files in the dataset are stored as *.wav with the following characteristics:

• 16 kHz sampling rate
• mono format
• audio codec = pcm_s16le

Runtime of main.py script

To run everything in the script it took 2.44s