5 ms·
How can they be 6 DOF if they're recorded with two cameras?
by stavros 3y ago
How can they be 6 DOF if they're recorded with two cameras?
- andybak 3y agoAI magic. (They aren't 6dof but it is possible)
- starcraft2wol 3y agoAI is not magic, and still constrained by math. If you need 3 variables and only have 2, AI isn't going to do it.
- andybak 3y agoUnless we're talking about different things - it already exists. There's two parts to it - estimating a depth map ("monocular depth estimation") and inpainting the occluded parts.
- Torkel 3y agoThey have a lidar. So they have very accurate depth. This gives some limited possibility at extrapolating and showing things from different viewing angles. I think this is also how they can give a good stereo experience from cameras that are closer than the average human IPD.
- j4yav 3y agoIt would have to be AI generated, right? The cameras can't see behind what they can't see behind.
- Torkel 3y agoMaybe, but not necessarily. Good oldfashioned computer vision can go a long way. When it comes to photography "AI" is a spice best used in moderation, cf eg https://www.theverge.com/2023/12/2/23985299/iphone-bridal-photo-three-poses-explanation-panorama-photoshop-generative-ai https://www.theverge.com/2023/12/2/23985299/iphone-bridal-ph...
- alexdbird 3y agoOnly for objects up to 5 metres away, though those might be where the most is gained from accuracy.
- Torkel 3y agoHuman depth perception only works up to ~7 meters so should be ok.
- ibrahimsow1 3y agoI like to imagine with 3D gaussian splatting of the videos. Not compute possible yet.
- kobalsky 3y agosomething like this? https://guanjunwu.github.io/4dgs/ https://guanjunwu.github.io/4dgs/