Zum Inhalt springen
Eye-tracking with AI

Look! Eye-tracking glasses

An affordable and reliable eye-tracker that works with your habitually worn eyewear. Eye-tracking glasses have raced each other in sub-degree accuracy. But once you take them outdoors, with natural illumination and reflections, there’s not so much left of it. We focus on robustness and produce a reliable sample (without compromises on accuracy).

Consequently, we left traditional eye-tracking with pupil detection, eye modelling and personalized gaze estimation behind (it’s just a single configuration setting away though). Instead, we apply a state-of-the-art deep learning pipeline that offers improved reliability under challenging conditions. Our networks learns a latent eye model and is able to infer a gaze direction that can still be calibrated just as a regular eye-tracker.

Simplicity

Easy to use, yet so powerful

All you need is 2-buttons: start/stop calibration and start/stop recording. Eye-tracking has never been easier.

We replaced expensive high precision manufacturing steps with sophisticated software. That enables us not only to produce affordable devices but also provides robustness to device slippage. You can even take the glasses off and put them back on. We produce a flexible frame that adjusts to many different head shapes. The deformation of the measurement device is then compensated in our software.

Get your new eye-tracker now!

Technical specifications

Download the datasheet with detailled information (pdf)

Eye-Tracking

  • Eye tracking technique: Appearance based gaze estimation
  • Binocular eye tracking: both eyes can be calibrated separately, compensating for misalignments.
  • Sampling rate: 30Hz (60Hz available on request)
  • Calibration procedure: 1-point and smooth pursuit
  • Pupil measurement: Pupil size measures are currently not available.
  • Accuracy: usually ~1° in real-world usage. Much better in the lab (see technical document for details)

Physical

  • Material: PA12
  • Scene camera format and resolution: 640 x 480 px, FOV of 120° horizontally, 70° vertically (that’s really large!)
  • Frame dimensions: 17 x 14 x 9 cm
  • Cable length: 1.4 m

Interface

  • API: Easy to use HTTP/JSON control interface. Real-time gaze data access via TCP. Real-time video access via WebRTC and MJPG. CSV export.

Get ready to record interactions between subjects with multiple devices at the same time! Just buy two or three and still save money.

Ever spent several days recording in the lab, a supermarket or a museum to get a representative sample size? Measuring subjects one by one is a waste of time. Acquire data faster, in parallel!

Ever moved outdoors with your eye-tracker? Ours keeps working.

Need a lot of devices, but only for a short time? Just rent a bunch.

FAQ

OS / hardware requirements

Currently only Windows is supported. We can run with Parallels on a Mac. There is also a special software version that runs on a Raspberry Pi. If you feel lucky and want to try a Linux build, we’d be happy to help out.

Get me started!

Grab our software over here! There is also a Quick Start Guide available in the Downloads section.
Connect your device via USB2.0.
Calibrate it with the calibration marker and you are good to go. Happy tracking!

Help!

We’ve got you covered. Write an Email to info@eyetracking-research.de or call us! +49 (0)152 56150263

From science to innovation

Our Story

Our office is located in Tübingen, Germany. We produce locally, assemble in-house and ship worldwide.  

In 2012, we tried using eye-tracking technology outside of the lab for the first time – in a driving school car. We wanted to know if people with visual field defects (blind spots) exhibit compensatory gaze patterns. And the technology failed us miserably. It just wasn’t designed for usage out of the lab. Measuring through glasses, with make-up and under real lighting conditions was simply too much for the systems of that time. We were able to throw a good part of the recordings in the bin. Frustrating. And yet a great starting point for exciting research.

We have laid the foundations with more than 40 scientific publications. Now we are bringing a new generation of devices back to the streets.

Dr. Thomas Kübler

Co-Founder and CEO

Eye tracking outside of the lab, e.g. in the car, is quite complex from an engineering perspective. At the Chair of Media Informatics / Human Computer Interaction at the University of Tübingen and the Chair of Vision Research at the University of Aalen, we developed methods to make eye tracking better.

Elena Wagner

Co-Founder and CMO

Elena is an expert in eye-tracking and its applications but beyond that also in marketing. She makes all those colorful ADs that might have brought you here.

Visit us

Sand 14
72076 Tübingen
Germany

Call us

Call us now at (+49) 152-562 502 63

WordPress Cookie Plugin von Real Cookie Banner