EyeScout: Active Eye Tracking for Position and Movement Independent Gaze Interaction with Large ...
summary
EyeScout mounts a commercial eye tracker on a motorized rail and uses a Kinect body tracker plus an Arduino-driven stepper to move the sensor to each user, enabling two modes: walk then interact and walk and interact. A 26-participant study shows gaze kick-off time of 3.5 seconds — a 62% faster start than prior guidance approaches — extends a single sweet spot into a sweet line, but currently supports only one user and has higher error and slower selection at faster walking speeds.
AI-written summary · about
chapters (8)
- 0:00Introduction and setup
- 1:22Large display challenges
- 3:05System design
- 6:23Technical implementation
- 8:56User study and results
- 12:00Future work and limitations
- 13:57Conclusion
- 15:38Q&A session
related talks
17mEyeGrip: Detecting Targets in a Series of Uni-directional Moving Objects Using Optokinetic ...
20mBuilding a Personalized, Auto-Calibrating Eye Tracker from User Interactions
19mGazeProjector: Accurate Gaze Estimation and Seamless Gaze Interaction Across ...
16mEye-Trace: Segmentation of Volumetric Microscopy Images with Eyegaze
48mEye Tracking (a Deep Learning Approach)
17mFast and Unassisted Eye Tracker Calibration for Gaze-Based Pervasive Human-Computer Interaction
09mUsing Gaze Behavior and Head Orientation for Implicit Identification in Virtual Reality
21mSelf-Calibrating Head-Mounted Eye Trackers Using Egocentric Visual Saliency