Master

Master's thesis for unsupervised online speaker diarization

To start

$ sudo apt-get install libasound-dev portaudio19-dev libportaudio2 libportaudiocpp0 ffmpeg python-pyaudio python3-pyaudio
$ pip3 install git+https://github.com/Desklop/WebRTCVAD_Wrapper
$ pip3 install lxml SoundFile pyannote.core pyannote.audio pyannote.metrics scipy pydub numpy librosa sounddevice webrtcvad PyAudio
$ pip3 install --no-cache-dir Resemblyzer
$ mkdir results
$ mkdir chunks

To prepare transcripts file for AMI recording

• select recording transcripts (ami_corpus/ami_public_manual/words, only ES2002a is left, everything is available at http://groups.inf.ed.ac.uk/ami/AMICorpusAnnotations/ami_public_manual_1.6.2.zip)
• set correct variables in data_preparation/__init__.py
• run script to make transcripts.json

$ python3 data_preparation/__init__.py

To run online diarization

$ python3 audio_recording/online.py

To run file diarization

$ python3 audio_recording/file.py