I'm working on training an image set that has some occasional obstructions of the face, so now that I have a more capable system I've been playing around with GAN to see if that helps in masking out facial occlusions. Other discussions mention that GAN isn't quite ready (or may be broken). That was a few months ago so I'm just curious if there's been any progress on that front?
Also, another discussion mentioned training without perceptual loss for 10k iterations, then training with perceptual loss for another 10k iterations, then to "turn on mask refinement." I don't see an option for mask refinement anywhere in the training. Is there a way to do this currently, or is it something that's been removed?
I'm working on training an image set that has some occasional obstructions of the face, so now that I have a more capable system I've been playing around with GAN to see if that helps in masking out facial occlusions. Other discussions mention that GAN isn't quite ready (or may be broken). That was a few months ago so I'm just curious if there's been any progress on that front?
Also, another discussion mentioned training without perceptual loss for 10k iterations, then training with perceptual loss for another 10k iterations, then to "turn on mask refinement." I don't see an option for mask refinement anywhere in the training. Is there a way to do this currently, or is it something that's been removed?