mdnahas

It is now common to have multiple people using their smartphones to video the same event. I'm thinking Pretti and Good's killings. I've heard of Gaussian Splattering, which constructs a 3D scene from multiple cameras. Is it useful for these analyzing these events? And, if so, can someone build an easy-to-use open source tool?

My speculation is that it would be useful to: (1) synchronize video, (2) get more detail than a single camera can get, (3) track objects (like Pretti's gun) that are seen by multiple cameras, and (4) identify AI generated video.

The last is most important to me. There is a danger of AI generated or modified video of an event. It seems possible to me that Gaussian Splattering from N videos will be able to detect if the N+1 video is consistent or inconsistent with the scene.

Is this possible?