Does anyone understand the business angle for meta here with these models? I still don’t understand why their research division exists and how it relates to their core business. I’m a huge admirer of their work but don’t understand the why.
Neat. It's mentioned on Facebook's page, but here Google's version of point tracking: <a href="https://deepmind-tapir.github.io" rel="nofollow noreferrer">https://deepmind-tapir.github.io</a> which is Apache-2.0 licensed.
I think Meta's goals are becoming clearer: they want to make VR <i>unbeliavable</i>. Judging by this and by SAM, they want an AI system than can understand the world around itself in <i>real time</i>.
I wonder how research works inside Product companies.<p>As an engineer working in a Product company, my focus switches between priorities set by PM and quarterly re-adjustments of the goals/strategies.<p>Can't imagine same thing can be done in research:<p><pre><code> * Hey, when are we releasing model for tracking any points?
* Can you estimate how long it takes to fix the issue you found in tracking accuracy?</code></pre>
I wonder how this compares to the motion estimation algorithms in the x264 and x265 video codecs. If it's better, then it can be used to increase video compression, by using it at the motion estimation stage for these codecs.
15 years ago on single CPU core laptop <a href="https://www.robots.ox.ac.uk/~gk/PTAM/" rel="nofollow noreferrer">https://www.robots.ox.ac.uk/~gk/PTAM/</a><p>Parallel Tracking and Mapping for Small AR Workspaces (PTAM) <a href="https://www.youtube.com/watch?v=Y9HMn6bd-v8">https://www.youtube.com/watch?v=Y9HMn6bd-v8</a><p>No ML, pure computer vision in C.
One present modern challenge I’ve noticed is reverse video searching. There are no good platforms for this like there are reverse image. I wonder if this ability to quantize videos would allow you to build more efficient indices of videos that you could check against from some input.
Oh wow, this will be the end of complicated motion capture rigs.
Indie developers can do motion capture for their own complex 3D characters in the comfort of a home. Good.