Risk Assessment For Installation of Drainage Pipes.pdf
A probabilistic approach to online eye gaze tracking without explicit personal calibration
1. A PROBABILISTIC APPROACH TO ONLINE EYE GAZE TRACKING WITHOUT
EXPLICIT PERSONAL CALIBRATION
ABSTRACT
Existing eye gaze tracking systems typically requirean explicit personal calibration
process in order to estimate certainperson-specific eye parameters. For natural human computer
interaction, such a personal calibration is often inconvenient andunnatural. In this paper, we
propose a new probabilistic eyegaze tracking system without explicit personal calibration.
Unlikethe conventional eye gaze tracking methods, which estimate theeye parameter
deterministically using known gaze points, ourapproach estimates the probability distributions of
the eye parameterand eye gaze. Using an incremental learning framework,the subject does not
need personal calibration before using thesystem. His/her eye parameter estimation and gaze
estimationcan be improved gradually when he/she is naturally interactingwith the system. The
experimental result shows that the proposedsystem can achieve <3° accuracy for different people
withoutexplicit personal calibration.