How to affine the image?

[alfredtofu]

As mentioned in your paper, you use adience_align to align the faces. So can you briefly explain the alignment setting. Just like how big the face in the image.

[GilLevi]

Hi,

Thank you for your interest in our work.

Alignment is done by detecting about 68 facial landmark and applying an affine transformation to a predefined set of locations for those 68 landmarks.

I didn't understand what you mean by the size of the face in the image. Can you please elaborate?

Best,
Gil

[ma1onso]

@GilLevi is needed alignment the images with https://github.com/eranid/adience_align before processing with your model?
Thanks

[GilLevi]

Hi @alonsoenrique , if you want to process new images (not from the AdienceFaces set which are already aligned) then yes, you will get better results if you will align them first. You will get a bit worse results without alignemnt, but I can't say in advance how much the results would be worse.