When AI and dance dads collide (spoiler alert, we’re a dance family). Here are six examples of Meta’s Segment Anything Model (also known as object tracking) applied to dance. Now, I can finally send relatives videos and say, “The two neon dancers are ours!”… LOL.
1) Tracking Rori’s tumbling pass at a basketball game.
2) Following Chloe and Rori practicing “Next To You” at X Squad Dancers
3) Tracking Chloe and Lexi in “Lost at Sea” at Imagine Dance Challenge
4) Isolating the same two dancers
5) Text integration on a tumbling pass (super cool for promo videos)
6) Burst effect test on tumbling pass.
Here’s another example using the Delmarva Sports Network’s plays of the week.
This is a free AI model from Meta. Object segmentation is fun, but it’s also critical for autonomous vehicles, robots, and, as you can imagine, for policing, security, warfare, and—of course—sports. If you think Moneyball or sign stealing is complicated, just wait until teams start analyzing poker tells, strengths, and weaknesses for every player. Every player, every move, every sport – tracked, logged, and run through predictive models.
- Compare this: https://showlab.github.io/magicanimate/
- To what Connor Stallions does (below)
- Combine it with segmentation (above)
Leave a Reply