The rectangles used for face recovery don’t really work. They often miss face elements, like nose tips, ears, necks, and more. They also overlap areas of the background, hairlines, and clothing, which I must manually correct. Wouldn’t it be better to allow me to paint in a selection, as with sharp and denoise? That way I can tell PhotoAI what I need to recover. I might, for instance, like to include hair (it does a great job with hair), necks, ears, and chests while avoiding backgrounds, clothing, etc.
