4 ms·
The difference here is that they (FB) are able to create 3D model based on only one photo. If you have multiple photos from different angles, you can just use g
by burntoutfire 7y ago
The difference here is that they (FB) are able to create 3D model based on only one photo. If you have multiple photos from different angles, you can just use geometric relationships to figure out where the points lie in 3D space. However, with only one picture, you need to "guess" (using ML), as there's not enough data for proper 3D reconstruction.