Stelfie the Time Traveller is a social media user who has been creating a time-hopping travelog with the usage of generative AI all through December. He used Stable Diffusion and fine-tuning. He is apparently an anonymous artist and created a fictional as well as photorealistic character which he can add into historical photographs set in different eras and time zones. Be it pictures from the time of the dinosaurs or ancient Egypt, all of it can be covered.
Stable Diffusion is a deep learning image synthesis model that offers people the ability to make fictional scenes using prompts that are text descriptions. Along with this Dreambooth technique can be used so that people can add their own subject or character inside the scenes that are generated by Stable Diffusion. Real people can also be added to fictional environments using the same.
As of now, Stelfie has captured selfies based in historical setting during the ice age wherein he is being chased by a wooly mammoth, in ancient greece with the Trojan Horse, in ancient Egypt during the construction of the pyramids, a picture wherein he is hanging out with Leonardo da Vinci while he is woking on painting the Mona Lisa, one while sailing with Christopher Columbus and one in the old West while running away from a Tyrannosaurus rex.
The 41 year old man behind Stelfie used social media platforms like Reddit, Twitter and Instagram to share his creations. For instance he wrote with the picture with Columbus that read “Log #3 – Columbus ship, 1492 DC. “Laaaaand” Now, I don’t wanna brag but… After months of sailing Columbus was lost in the Oceans and well let’s say I may have kinda suggested the correct direction…”
He also gave some details on the process he uses to make these images that can be found on the comment section of Reddit in one of which he writes: a combination of Stable Diffusion 1.5, a custom AI model for the landscape, and a custom AI model trained on the Stelfie face, which is apparently a fictional person created using Character Creator.
ALSO READ | Five trends set to shape digital transformation in 2023
ALSO READ | After ChatGPT, OpenAI launches open-source tool Point-E that converts text into 3D image
