Machine learning can detect baseball signs?

Experienced youtuber Mark Rober, who worked as an engineer at NASA, developed a fun application this time using machine learning to decipher baseball signatures. In baseball, coaches, coaches, and catchers use specific, complex gestures to direct players to action. The app detects these gestures and predicts behavior.

The application we created this time distributes the touch points when signing, from A to H, and when you tap this sequence, it identifies which direction it is. Of course, some of the complex signs deliberately hide meaningless things, but this app learns everything with just three inputs, so you can see even this.

He heard about 50 players and coaches for this project. For example, if you touch the hat, you steal, and all the rest is bait, or if you don’t touch the hat, it’s all bait. All the rest is a mix of meaningless movements.