Robots and embodied-AI systems learn from human demonstration. Video captures the action. Attention captures the intent behind it: which object was chosen, which hazard was checked, what guided the decision.
ETVision produces synchronized, object-grounded attention data: 180 Hz binocular gaze, fixations, pupil dynamics, and gaze on named objects from a model you can train on your own classes.
Not a dot on a video. Gaze tied to the objects that matter, via an on-platform model you train in ETAnalysis.
Everything runs on the platform. No public-cloud routing. Your scenes and your dataset never leave your control.
Tell us about your task and objects. We will show ETVision capturing the attention layer.
Talk to an engineer