I think it'd be better if we avoided including the extra image processing techniques here (at least for the moment). It's better to feed the neural nets with the extact same preprocessing they had for training. However, I'll take note as they're interesting preprocessing techniques for data augmentation :)
As for the weighted face detection part, I think it's a very nice addition! Several people requested having 2+ faces on frame. This should fix it nicely :)
If you don't mind, please, revert (or stash) the changes on the ImageProcessor and set the target branch to dev. I'll merge this addition and test it as soon as I can.
Hi!
I think it'd be better if we avoided including the extra image processing techniques here (at least for the moment). It's better to feed the neural nets with the extact same preprocessing they had for training. However, I'll take note as they're interesting preprocessing techniques for data augmentation :)
As for the weighted face detection part, I think it's a very nice addition! Several people requested having 2+ faces on frame. This should fix it nicely :)
If you don't mind, please, revert (or stash) the changes on the
ImageProcessor
and set the target branch todev
. I'll merge this addition and test it as soon as I can.Thank you very much!