Thanks for your demo, you said if I want to analysis every person's gender and age at one image, I should use a face detection network to extract every person's face firstly, so I can use mtcnn to detect face,but should I preprocess the face image before send it to your network? I mean， should I align the face image, or not? extend the face rectangle, or not? I think I should extend the face rectangle to include more parts of hair,ear and jaw, is it right?
Thanks for your demo, you said if I want to analysis every person's gender and age at one image, I should use a face detection network to extract every person's face firstly, so I can use mtcnn to detect face,but should I preprocess the face image before send it to your network? I mean， should I align the face image, or not? extend the face rectangle, or not? I think I should extend the face rectangle to include more parts of hair,ear and jaw, is it right?