Hey, I find myself wondering whether the LIDAR sensor (on devices that have it) is utilised during the process of Gaussian splatting?
If it isn’t, is there value to performing a normal mesh based LIDAR scan in a splat-like way and then reprocessing the raw data afterwards as a splat?
Might seem weird but its’ a question that keeps bothering me so why not just ask. Also if there are any specific technical tips you guys have beyond the fairly surface level content that you can share I’d be extremely happy! It’s a very impressive feat having the real-time on device scan thing even if from a pure user satisfaction perspective, so I’ve no doubt it’s a heavily guarded piece of intellectual property, but it would be great if you could share what you can!
P.S always up for a testflight invite if a private/semi public dev pipeline exists.