GenderRecognition

Caffe Model to detect Gender

models folder contains

Models can be downloaded from here

• age_gender_mean.t7
• haarcascade_frontalface_default.xml
• deploy_gender.prototxt
• gender_net.caffemodel
• mean.binaryproto

How to use

• Caffe Installation here
• Torch Installation with Luahere
• Other installations (Run 'luarocks install dpnn', 'pip install lutorpy')
• Download the trained model zip file from the link above and unzip the file
• Put all the model files in folder 'models'
• keep the codes and 'models' in the same location
• Make sure Camera is connected as USBID is 0(default)
• To check different implementation, run the following codes
  • python gender_demo_lua.py : Python wrapper to run Lua code, fastest code
  • th gender_demo_torch.lua : Lua code to detect gender, best accuracy
  • python gender_demo_caffe.py : Caffe code, slower

Lua implementation source :

https://github.com/szagoruyko/torch-opencv-demos/blob/master/age_gender/demo.lua

Caffe implementation source:

http://nbviewer.jupyter.org/url/www.openu.ac.il/home/hassner/projects/cnn_agegender/cnn_age_gender_demo.ipynb

References

• https://gist.github.com/GilLevi/c9e99062283c719c03de

• http://www.openu.ac.il/home/hassner/projects/cnn_agegender/

• http://www.openu.ac.il/home/hassner/projects/cnn_agegender/cnn_age_gender_models_and_data.0.0.2.zip

• As far as speed is concerned, you can check a discussion over here. GilLevi/AgeGenderDeepLearning#6

• The subtraction from the mean is a way of normalization, please find a detailed answer here http://stats.stackexchange.com/a/220970