This space is a collection of small explorations, half-finished ideas, and things I’m trying just to see what happens. Some are rough, some might turn into something bigger, and some will stay exactly as they are. Think of it as a public notebook — open, evolving, and experimental.
I finally had a look at Apples: Sharp Monocular View Synthesis:
machinelearning.apple.com/research/sharp-monocular-view
With this algorithm you can turn a single photograph into a 3d gaussuan splats pointcloud.
I tested it with some images from my trip to Columbia and i think it can be used as a modern kind of Ken Burns effect quite effectively:
I´ts just a personal project, still work in progress to celebrate this great country, people and nature:
temporary music by the great Imogen Heap – Hide and Seek
I created a workflow with the help of Ai to turn a single image into a 3d character, which i can rig and animate with motion capture data: it turned out rather well and was fascinating how quick such things can be done, at least for a kind of background character. i Also added a quick cloth sim for the lower part of the dress, that was done in a conventional 3d way:
For the project “Outsider Freud”, we had some scenes inside and outside a tramway, so i went to the Strasssenbahnmuseum and did a 3d scan and Gaussian splatting scene of the inside of a tramway from that period.
Here is a Link to see the gaussian and navigate through it: (it takes some time to load, ~80 MB)
Recent Comments