> I wonder if the "deep learning" is getting any hints from the bayer patterns that may be detectable even in the desaturated image.
I'd say it's highly unlikely, since these images are downsampled to 224x224 pixels. That would average out any residual Bayer pattern (which is pretty hard to detect in the first place).
I'd say it's highly unlikely, since these images are downsampled to 224x224 pixels. That would average out any residual Bayer pattern (which is pretty hard to detect in the first place).