# Pre-AttentiveGaze: gaze-based authentication dataset with momentary visual interactions

## TL;DR

The paper introduces a dataset for gaze-based authentication using rapid eye movements in response to visual stimuli.

## Contribution

The novel contribution is a dataset capturing eye movements during pre-attentive processing for authentication purposes.

## Key findings

- A dataset of 76,840 eye movement samples was collected from 34 participants.
- Gaze features were extracted and validated using machine learning models.
- The dataset shows potential for gaze-based authentication using pre-attentive processing.

## Abstract

This manuscript presents a Pre-AttentiveGaze dataset. One of the defining characteristics of gaze-based authentication is the necessity for a rapid response. In this study, we constructed a dataset for identifying individuals through eye movements by inducing “pre-attentive processing” in response to a given gaze stimulus in a very short time. A total of 76,840 eye movement samples were collected from 34 participants across five sessions. From the dataset, we extracted the gaze features proposed in previous studies, pre-processed them, and validated the dataset by applying machine learning models. This study demonstrates the efficacy of the dataset and illustrates its potential for use in gaze-based authentication of visual stimuli that elicit pre-attentive processing.