Personality-Driven Gaze Animation with Conditional Generative Adversarial Networks

TL;DR

This paper introduces a GAN-based method to generate personalized gaze behaviors for virtual agents based on Big-Five personality traits, enabling more realistic and personality-aligned gaze animations in virtual environments.

Contribution

It presents a novel generative adversarial network model that synthesizes gaze behavior conditioned on personality traits, trained on eye-tracking data from real participants.

Findings

Successfully generated gaze patterns aligned with personality traits.

Enabled realistic gaze animation in a game engine.

Demonstrated the model's ability to produce diverse gaze behaviors.

Abstract

We present a generative adversarial learning approach to synthesize gaze behavior of a given personality. We train the model using an existing data set that comprises eye-tracking data and personality traits of 42 participants performing an everyday task. Given the values of Big-Five personality traits (openness, conscientiousness, extroversion, agreeableness, and neuroticism), our model generates time series data consisting of gaze target, blinking times, and pupil dimensions. We use the generated data to synthesize the gaze motion of virtual agents on a game engine.