METAL for iPhone

Read AI news in the METAL app.

Download METAL and discover fresh AI stories every day.

Download on the App Store

For iPhone · Free download

Search for METAL AI Magazine in the App Store on your iPhone.

METAL

AI GlossaryㅅTechnical words in the news

hand tracking model

AI technology that finds the position of a hand and the shape of its fingers in real time from camera footage

In plain words

A hand tracking model is technology that lets a computer figure out where a hand is and what shape the fingers are making, just by looking at camera footage. It's similar to how a camera can automatically detect a face and draw a box around it—except here the target is a hand, and the system pinpoints the position of each individual finger joint.

In the past, doing this required sensor-equipped gloves, special cameras, or other dedicated hardware. Now, software can calculate hand movement on screen using nothing but an ordinary webcam already built into a laptop or smartphone. Open your hand and it registers as "open"; close it and it registers as "closed"—all extracted as coordinate data.

This extracted hand-position data can be used directly by other programs. By mapping how open a hand is to volume, and the hand's height to pitch, you can build an instrument played purely through hand gestures. The same principle is used in game controls and in recognizing hand gestures during video calls.

How it shows up in the news

The article introduces the browser-based instrument "Air Theremin," explaining that "we've reached the point where a browser's webcam API and a hand tracking model alone can mimic the same experience." It's used to convey that while hand recognition once required dedicated sensors or hardware, it can now be done with just a webcam and software, with no extra equipment needed.

Try it yourself

Open a hand-tracking-based web service or app on a laptop or smartphone with a webcam, then open and close your hand or move it side to side within the camera frame to check the response speed and accuracy for yourself. Try moving your hand out of frame to see if tracking cuts off instantly, and test how well it works in dim lighting—doing this lets you feel the real limits of a hand tracking model firsthand.

See also

Stories using this term

Browse every entry