I never really got the motivation for this. Please correct me if I'm wrong but it seems like if I am creating a neural net to recognise if someone is smiling, I should use all the information in the picture, including race and gender.
No comments yet.