Seems plausible. However I think better proof that a particular method was used to generate an image is to identify a way in which the generated image is incorrect and show that the images you hypothesize might have been generated in that way have that defect. The artifacts of the method are likely to be more specific than the fact that the results look good (because all methods are trying to make the results look good).
This, however, is nonsense:
"The way Vermeer painted this wall is consistent with a photograph. It is not consistent with human vision. If you were standing in the room that Vermeer painted, you would see that wall as a pretty even shade of off-white. "
I have only the barest of artistic training, and a bit of graphics programming experience and I can very clearly see that the brightness of the wall varies with proximity to the window which lets in the light and various obstruction (the fact that he is pointing at the picture and assuming we can all see the variation that the artist cannot possibly see makes the claim particularly absurd). I would assume that if I spent many years studying and practicing to be a realistic painter I would be vastly better at that. Even if some eye processing was muting the differences I'm sure at least a tiny minority of painters could learn to compensate with years of practice. One of the fundamental skills in painting and drawing is to learn to see the image in a less abstracted way.
Also the photograph shown clearly has much more brightness variation than either painting, particularly on the bit of ornate furniture the man is leaning on.