Object tracking explained in 3Blue1Brown style. Learn about the tracking loop, kalman filters, why occlusion breaks and one way to fix it. Try it out

14 box is huge tf

You are so crack 😍😍😍

Thanks!

Simulated IMU data with heading and gate analyzed. That’s what I would do.

But the core learning from ai is ... Desisit from designing these case specific algorithms.. Just good data and learning for these and 1000 other cases. . delving in single case appears smart but doesn't scale.

i dont think the bitter lesson applies to everything. We still need ligthweight algorithms for a lot of use cases. A big model wont run 30 fps on an edge device, this does

Indeed a good point ... A single fast algorithm has its merit .. the con side is disconnected from a generic web of intelligence ... Breaks at too many edge cases ... Maybe use a very small nn over there ... Many layers are possible..

What type of science or mathematics would this be .

Dude was never seen again!

Just say she's fat

Great explainer - Squash has been my occlusion test; two players in one box, crossing on every shot. Here's how it looks so far