Learning to Personalize in Appearance-Based Gaze Tracking

TL;DR

This paper introduces SPAZE, a low-dimensional latent space model for appearance-based gaze tracking that adapts to personal variations with minimal calibration, achieving state-of-the-art accuracy and enabling head pose-free tracking.

Contribution

The paper proposes SPAZE, a novel low-dimensional latent parameter model that effectively captures personal variations in gaze tracking, improving calibration efficiency and accuracy.

Findings

SPAZE achieves 2.70° error with 9 calibration samples.

Personal variations are modeled as a 3D latent space per eye.

Head pose-free gaze tracking is demonstrated.

Abstract

Personal variations severely limit the performance of appearance-based gaze tracking. Adapting to these variations using standard neural network model adaptation methods is difficult. The problems range from overfitting, due to small amounts of training data, to underfitting, due to restrictive model architectures. We tackle these problems by introducing the SPatial Adaptive GaZe Estimator (SPAZE). By modeling personal variations as a low-dimensional latent parameter space, SPAZE provides just enough adaptability to capture the range of personal variations without being prone to overfitting. Calibrating SPAZE for a new person reduces to solving a small optimization problem. SPAZE achieves an error of 2.70 degrees with 9 calibration samples on MPIIGaze, improving on the state-of-the-art by 14 %. We contribute to gaze tracking research by empirically showing that personal variations are well-modeled as a 3-dimensional latent parameter space for each eye. We show that this low-dimensionality is expected by examining model-based approaches to gaze tracking. We also show that accurate head pose-free gaze tracking is possible.