Gesture Control for Linux With a Standard Webcam

Here is the short answer. Linux has no built-in camera gesture control, and most of what you find online is a Python script built on an open source hand tracking library. AirTouch is a finished alternative. It runs on Linux, reads hand gestures through a standard RGB camera, and converts them into the mouse, keyboard, and touch input your applications already understand. There is no SDK and no change to the software you are controlling.

Why does gesture control on Linux matter?

Look behind almost any screen that is not a personal computer and you will find Linux. Kiosks, digital signage players, museum exhibits, industrial panels, and many medical imaging workstations run it. Those are the same places where touching the screen is the problem: shared surfaces, gloved hands, displays behind glass, screens mounted out of reach.

So the platform with the most need has had the fewest options. Many commercial gesture products focus on Windows.

What are the options for hand tracking on Linux?

• Build it yourself. Open source libraries such as MediaPipe give you hand landmarks in Python. You write the gesture logic, the cursor smoothing, and the input injection. Good for learning and for custom apps. See MediaPipe hand tracking: build it or buy it.

• Sensor-based SDKs. Depth cameras and dedicated hand sensors often include Linux SDKs. You get data, and your application has to be written to use it.

• AirTouch. Finished gesture control and hand tracking software that works at the operating system level, so any application responds as if a mouse, keyboard, or touchscreen were attached.

How does AirTouch work on Linux?

The interaction is the same as on Windows. Raise an open palm to take the cursor. Pinch your index finger and thumb to click, and hold the pinch to drag. Fifteen gestures in total can be mapped to keyboard shortcuts and mouse actions, and mappings save per profile. Touch mode sends touch events instead of mouse events for interfaces designed for a touchscreen.

Gesture inference runs on the machine, and camera video is processed locally. The gesture list and mapping guide applies on every platform.

What do you need?

• A camera. A standard USB webcam that delivers 20 frames per second or more. Below 15, the cursor will feel laggy. The camera guide has recommendations.

• A reasonable CPU. Our baseline is a 10th generation Intel Core i3 with 8GB of memory.

• An internet connection. AirTouch needs one to install and to run.

• Your distribution details. Linux environments vary. Tell us your distribution and display setup and we will confirm the right build for it.

Where is it used?

• Kiosks and wayfinding. A touchless layer over the kiosk application you already run. See the touchless kiosk.

• Digital signage. Proximity Trigger wakes content when someone walks up, and gestures let them browse. See gesture control for digital signage.

• Imaging workstations. Clinicians navigate viewers with gloved hands. See gesture control for medical environments.

• Your own desk. Flip through terminal tabs like a conductor, pause a video from across the room, or lock your screen with a gesture as you walk away.

Does it need a depth camera?

A standard RGB camera covers typical use. Dark rooms and precision spatial work are a different story. Through our partnership with Orbbec, AirTouch also takes advantage of 3D depth sensing and IR cameras when a deployment calls for them. The Orbbec page explains when that is worth it.

Where to start

Start a free trial, or contact us with your distribution and use case. Product teams building on embedded Linux can ask about OEM licensing on the OEM page.

Related: gesture control for Windows, control your computer with hand gestures and a webcam, and gesture control and hand tracking software.