In the context of the project Mosul, we need to build models from a set of pictures which can be damaged, old etc..
The pictures of the museum in good state can be quite old or damages, and we have also modern ones (more or less let say).
I don't have experience in this field (yet), my purpose here is more about curiosity if someone want to try and give a feed back.
The experience can be:
Using another item, already 3D scaned here (not necessary heritage related) and took picture in black and white + damage the pictures by image processing (add noise), try to 3D model it again with a mixt of good and bad picture, and adjust to get the original 3D model (only with good pictures).
In this way we can assess a method to deal with old picture among good quality one. We can repeat some method for controlled sample in order to produce better results for the project Mosul but also for others where we face the need of old pictures.