Gender Detection

This project is to test performance of different classification method on Gender Detection problem

Test tool usage

example:

python GenderDetect.py [options] wave_file model_file
  where options may include:
  -m Model --determine which model you want to use, including GMM,MLP,KNN,DBN(currently, only GMM is available)
  -l window length --the length of the analysis window in seconds. Default is 0.025s (25 milliseconds)
  -s window step  -- the step between successive windows in seconds. Default is 0.01s (10 milliseconds)

Training document

Explanation for the process of training.

Document	Description
voxforgeDownload.sh	Download voxforge corpus from website
Corpus_prep.ipynb	Pick and formalize the audios for further training
MFCC_extract.ipynb	Extract AMFCC features from prepared corpus
GMM_classify.ipynb	Train GMM models for gender detection
base.py	Provides mfcc features for use in ASR applications
sigproc.py	routines for basic signal processing
GMM.pk	Trained GMM models
male_example.wav	a male audio
female_example.wav	a female audio