No disrespect to brummy-lad in tidying up my photo in post #36 as the details brought out by AI are very interesting and helpful.
AI hasn't (in this case) worked so well on the faces, more specifically the eyes and ears. (Have to admit, the original photo seems to have been more focussed on the background than on the people). But the AI version is almost like a photofit where the different parts have been pieced together, and the eyes and ears don't quite blend. Mine and my brother's eyes are too large and my dads aren't scowling enough - and believe me he could scowl. Most other features are good, such as hairline and hands.
I was surprised to see my dad's tartan socks were brought out, and they definitely look tartan in the original. Didn't have him down as a wearer of tartan !
On objects, the AI version seems to have misrepresented a flaw on the original photo for a cigarette, I think. He never smoked in his life. AI seems to have made some sense of the item to the right - a seaside bucket. It didn't do so well with the baby feeding bottle between the carrycot and the bag on the original - it seems to have been AI morphed into a strap of some kind for the bag.
But like I say, it's helped to bring out many more details than in the original and am grateful for that as the original photo was only about 2" x 1.5" in size.
Perhaps AI has some way to go before it can more accurately represent detailed facial features and human expression ?