I was introduced to the power of Stable Diffusion a few weeks ago when I saw works of art literally created out of thin air by simply “describing” what you wanted the content of the image to contain. I explored the power of this a little more and discovered that it was also able to create a likeness of anyone through the use of “embeddings” where you produce a file containing the necessary parameters needed to steer the Stable Diffusion algorithm in the direction to the desired subject. This got me thinking …. oh … we can get Veronika out retirement and create a “metaverse” version. I searching around, I had found that someone on one of the AI forums had already created a basic Veronika model I could start with. After about 1.5 hours of loading the Stable Diffusion AI software on my server and then about another 2 days of playing around with settings, I was able to come up with these 2 almost photo realistic “new” AI created pictures of Veronika:
The resemblance if almost perfect. Of course there is still work to do to better refine the model but certainly a very good start.
If anyone is interested in learning more about Stable Diffusion and how it works you can consult the following Wikipedia link on the topic: https://en.m.wikipedia.org/wiki/Stable_Diffusion