Very cool product! And to think, in one of the many prior gaussian splatting threads someone declared there was no way anyone could build a business around the technique.
Incredible. I may be the only one in the dark, but until this moment I had no idea 3D printing at this high a fidelity was possible. It looks like a real bee.
Last I knew, the best 3D prints still looked like hardened play dough
Interesting how they voxelate the splat to print it. I am sure there are more efficient algorithms that allow you to make a mesh which then gets sliced before printing
But, if you made a mesh you wouldn’t be able to represent fuzzy objects. That’s one of the major advantages of both splats and this printing technique.
This is genuinely fucking cool. I don't think I've ever seen a 3d printing implementation like this before. I'm not a big fan of splats as a modeling method, but for something like this it's fantastic.
Once in a while I get inspired by a post on HN, but I know I don't have the energy to follow it through. Here's my idea for anyone who thinks they can do something with it:
Imagine you are holding in your hand a small glassy sphere. Initially it looks like random dots inside, but as you play with it in your hands, you get glimpses of images of loved ones.
To implement this, select a bunch of photos of people, generate pseudo-3d models of their faces with your AI of choice. Position them around a sphere in a 3D modelling software, and generate a whole bunch of renderings all around the scene. Now, feed those renderings into gaussian splat software.
The resulting 3D gaussian splat should include relatively high fidelity when the viewer is aligned with the supplied iamges, but more stochastic nonsense when misaligned. This should contribute to the feeling of being able to 'find' images while rotating the sphere around.
if you allowed for some slippage in precision, how many images could you fit in that are recognizable from various angles, with ~acceptable~ resolution? Sounds fun.
21 comments
[ 0.33 ms ] story [ 37.8 ms ] threadLast I knew, the best 3D prints still looked like hardened play dough
5mm x 5mm x 5mm (smallest I can select): $32.58
10mm x 10mm x 10mm: $36.58
20mm x 20mm x 20mm: $82.86
30mm x 30mm x 30mm: $143.60
50mm x 50mm x 50mm: $509.14
100mm x 100mm x 100mm: $3,028.58
150mm x 150mm x 150mm: $9,618.86
195mm x 195mm x 195mm: $20,641.32
(it says it's an estimate excluding tax and shipping)
Imagine you are holding in your hand a small glassy sphere. Initially it looks like random dots inside, but as you play with it in your hands, you get glimpses of images of loved ones.
To implement this, select a bunch of photos of people, generate pseudo-3d models of their faces with your AI of choice. Position them around a sphere in a 3D modelling software, and generate a whole bunch of renderings all around the scene. Now, feed those renderings into gaussian splat software.
The resulting 3D gaussian splat should include relatively high fidelity when the viewer is aligned with the supplied iamges, but more stochastic nonsense when misaligned. This should contribute to the feeling of being able to 'find' images while rotating the sphere around.
The cube is filled with splats. Each face reveals a different picture when viewed from a perpendicular angle
[0] https://www.3dpoint.art/