Prescription glasses, 3D gaze, local processing, accuracy, motion capture, and support for legacy ASL systems.
Yes. The ETVision frame is designed to sit over the participant's own glasses. Nothing has to be swapped out and no corrective lens has to be fitted to the tracker. Participants who need their glasses to do the task can keep wearing them, which matters when the task is flying, driving, reading an instrument or operating equipment.
Most eye trackers draw a dot on a scene video, which tells you where on the image the person was looking, not what they were looking at. ET3Space computes binocular vergence and combines it with head tracking, so gaze is reported in real-world room coordinates and mapped onto physical objects you define. You get the object, not the pixel position. 3D gaze in the real world explains how the world model behind it is built.
Post-hoc AOI tracking happens after the session. An area of interest is drawn around an object in the recorded video, and software follows it frame to frame. The result is where that area was, worked out recording by recording. Live AI object detection happens during the session. A trained model recognizes objects in the scene camera image and names the one gaze lands on, as it happens. In ETVision this is AI LMOT (Live Multiple Object Tracking). Per-object attention statistics build up while the task runs, and every participant's data arrives already labeled. The two connect in Argus software. ETAnalysis does post-hoc AOI tracking, and the moving areas of interest defined there can be trained into a custom model that runs live on ETVision.
No. Recording, gaze computation, object recognition and analysis all run locally, on your own machine. Nothing is uploaded to an external service and there is no account to sign into. Sites that cannot send data outside the building can run the system exactly as it ships.
ETVision samples each eye at 180 Hz and is specified at 0.5 degree accuracy. Tracking is binocular, so both eyes are measured independently rather than one eye being estimated from the other. Battery life is over four hours of continuous recording.
Auto calibration takes about five seconds, using automatic feature detection. Single-point and multi-point calibration are available when a study needs them. Calibration is what makes the accuracy figure true for the participant in front of you rather than inferred from a population average, which is why it is worth the five seconds.
Yes. Both eye videos are recorded at 180 frames per second, so any moment can be replayed frame by frame and the session recalibrated afterwards. Gaze and vergence are solved from both eyes by geometric computation, not estimated by a model, so the same eye video always gives the same answer. When a reviewer challenges a result, it can be traced back to the measurements.
Yes. ET3Space is built to take head position and orientation from an external tracking system and combine it with the eye data to produce gaze in room coordinates. This is how gaze gets mapped to instruments in a cockpit, surfaces in a vehicle mockup, or displays in a control room.
Yes. AI LMOT runs on the platform itself. A standard 80-object model is included with every ETVision, and the AIObjectTraining add-on for ETAnalysis trains custom models on the objects that matter in your environment. Because it is not routed through a public cloud service, the training data and the recorded sessions stay with you.
Yes, automatically. ETVision is fundamentally invariant to Parallax Error. By leveraging its knowledge of the 3D vergence angle, the system dynamically recalculates the scene-camera projection offset based on the user’s distance from the target. As a result, ETVision eliminates parallax error automatically, without requiring manual adjustment or calibration for different viewing distances.
Yes, automatically. Using the pupil, corneal reflections, and other eye features captured by the eye cameras, ETVision employs a 3D eye model to continuously determine the position and orientation of each eye. Because each eye is measured independently, the system can automatically compensate for changes caused by glasses movement.
Argus Science is the direct successor to Applied Science Laboratories. ASL was one of the first companies in eye tracking, with roots in research at MIT, and the line of work has run continuously for more than forty years. The people, the measurement methods, and the responsibility for the installed base carried over.
If you have an ASL eye tracker in a lab and need manuals, software or help getting it running, see Legacy and ASL support, or write to techsupport@ArgusScience.com.
Argus Science designs and builds its systems in Tyngsborough, Massachusetts. Hardware fabrication and software development are both done in the United States.
No. There is no annual fee to renew, in year one or year ten.
Compare them on the questions that decide a study. ETVision names the object a person looks at, live, with AI LMOT and an 80-object model included. It measures 3D gaze from both eyes, with depth, and with ET3Space places it on real objects in room coordinates. It is worn over the participant's own glasses. The software comes with the system, with no annual fee. Everything runs on your own system, with no cloud and no account. And each result can be reproduced and checked against the recorded eye video. The full list, with what to ask about any other system, is in the guide How to choose eye tracking glasses for research.
Yes, and it is free. The SDK ships with working samples in C++, C#, Python, LSL and MATLAB, so live gaze data can be pulled into your own application or synchronized with other instruments.
The wearable system underneath it. Both eyes measured independently at 180 Hz, which is what makes vergence a measurement rather than an estimate.
How gaze is placed on real objects in a room, and what a world model has to contain for that to work.
Eight questions that decide which wearable eye tracker fits a study, including the cases where the answer points elsewhere.