Eye pattern analysis in intelligent virtual agents

Abstract

This research discusses a new approach that uses patterns of a person’s eye motion to improve human computer interaction between intelligent virtual agents (IVA) and the people who interact with them. This approach adds new capability to current eye tracking interfaces that focus on eye fixations instead of patterns of motion. The analysis of eye motion on the pattern level can deliver three values to an eye tracking interface: speed, interaction reliability, and a more complete understanding of user attention. This research builds a system, called InVision, to demonstrate how the analysis of eye fixation at the pattern level can improve eye tracking interfaces for human computer interaction. An IVA is built using this eye pattern analysis technique.