/GenderDetect

Python wrapper around LIUM's gender detection toolkit.

Primary LanguageJavaGNU General Public License v3.0GPL-3.0

Gender Detection

Python wrapper around LIUM's gender detection toolkit.

Gender detection is the process of finding gender of speaker(s) from the audio file.

  • This wraps LIUM's gender detection tool.

  • LIUM uses pre-trained gmm models for detection which were trained with French voice corpus. So it is possible that gender output might be inaccurate.

  • This creates two files: .seg and .g.seg (learn more about them here)

  • You can delete these files after detection

  • Use when there is only one speaker in the audio file.


Note

The code(Python) is taken from the project voiceid.
Changes/improvements are made in order to work with minimum requirments.

Also, I modified the source code (java) of the SpeakerDiarization toolkit.
Source code is in the src folder.
See src/CHANGES.txt .

install

$ python setup.py install --user

Code

>> from gdetect import genderdetect as gd
>> gender = gd.identify_gender('./test.wav')
>> print gender
>> # will print 'F' or 'M'

Credits

VoiceId project