predict-facial-attractiveness

Using OpenCV and Dlib to predict facial attractiveness.

Download the shape_predictor_68_face_landmarks.dat from:
http://dlib.net/files/shape_predictor_68_face_landmarks.dat.bz2
and replace the file: data/shape_predictor_68_face_landmarks.dat

Then use your own path replace the root value in the *.py file,if there is a root.
Do not forget add a / (not a \) at the end of the path.
Just like this:root = 'E:/Github/predict-facial-attractiveness/'

After that,follow the steps below:
1.Run source/trainModel.py to generate the model which will be used to predict facial attractiveness later.
2.Run source/getLandmarks.py to get the landmarks of the faces detected in the image/test.jpg.
3.Run source/generateFeatures.py to generate the features,which will be as the input of the model we get before.
4.Run source/myPredict.py to predict facial attractiveness.

Or you can just run the source/run.py alone.