ETVision AI Object Detection and AI Model Training

Defined objects tracked live in ETVision.

Gaze and fixations on each object reported in real-time.

Objects can be "trained" for AI recognition with automated process using ETAnalysis Moving Area of Interest capability.

"Trained" objects will be auto-detected for all participants.

Real-time bar plots by object: Total Time, Fixation Time, Time to First Fixation, Pupil Size.

Argus Science launched the new AI Powered feature for ETVision: Live Multiple Object Tracking (AI LMOT). The Argus Science ETVision eyetracker can use an AI Model to recognize and detect multiple objects that appear in the head mounted scene camera image. The objects are detected and tracked in real time and the system flags periods of gaze engagement with each object in real-time.

Argus Science provides a standard AI model trained to recognize a set of common objects. An optional add-on to the Argus Science ETAnalysis application, called AIObjectTraining, enables users to easily create custom AI models for ETVision that will recognize objects specific to their requirements and environment.

In above screenshot ETVision has been loaded with a custom AI model trained to recognize the information display, instrument cluster, rearview mirror, left sideview mirror, and right sideview mirror. A recognition box tracks each of these objects in real-time and the system flags periods of gaze engagement with each object. Real time bar plots show cumulative fixation time on each object. Alternately, the bar plots can show Total Time, Number of Fixations, Average Fixation Duration, Time to First Fixation, or Pupil Size. Please check out our YouTube channel and view ETVision Live Drive with AI MOT video:

In above example, custom AI model has been trained to recognize golf balls. Please check out our YouTube channel and view ETVision-Golf Putting with AI MOT video: