We are very excited to announce our newest product - Pupil Invisible.
The first eye tracking device that truly looks and feels like a normal pair of glasses. No setup, no adjustments, no calibration.
Leveraging insights we gained working on Pupil, we have created a new gaze estimation pipeline powered by deep neural networks and matching eye tracking hardware.
Easy to use
Our approach does not rely on glints or even images of the pupil. There is no adjustment and no calibration required. Just put it on, and go.
Sunny outdoor environments
Traditional eye tracking approaches fail in bright sunlight. Pupil Invisible provides robust gaze estimation in any environment!
We managed to move the cameras completely outside of the field of view and fully integrate them into the frame. The first eye tracker that truly looks and feels like a pair of glasses.
Our algorithms perfectly and instantaneously compensate for the movement of the glasses, ensuring high quality eye tracking data at all times.
Pupil Invisible comes with an Android smartphone and companion app. It uses the companion phone for real-time neural network gaze estimation, real-time streaming over wifi, and recording.
Detachable scene camera
A wide angle camera records your full field of view. An integrated microphone can capture audio. The entire scene camera module can be quickly detached/re-attached via magnetic connection.
A high resolution Inertial Measurement Unit (IMU) is integrated into the frame.
You can easily pop in/out lenses with different prescriptions and shades.
Pupil Invisible closed beta program launches today!
Sign up for updates and early access to hardware.
We’ve spent 3 years working on Pupil Invisible, and are very excited to get it into your hands!
Pupil Invisible is a very new approach to mobile eye tracking. It does not replace the existing Pupil Headset. We see it as a complementary (and at this point) experimental new product!
We are looking for core users to explore new use cases and contribute to the final steps of product development in a closed beta program. The public (non beta program) release is scheduled for mid 2019.