Eye Tracking Dataset for Comprehensive Eye Tracking Feature Extraction

This repository contains eye tracking signals from the Tobii Pro Glasses 3 eye tracker. The study was conducted in a simulated driving environment using a motion-based driving simulator. The database contains signals from 26 healthy individuals. The task included Baseline, driving under normal conditions (Ride), and driving during fog (Fog). Sampling frequency is 100 Hz and the resolution is 1920 x 1080 pixels. Naming convention for files: ID_gazedata_output. For convenient test download and appropriate preview, we provided all signals in a .rar. Dataset contents Eyetracking-database.rar, data files, text format README.txt, metadata for data files, text format If you find these signals useful for your own research or teaching class, please cite both the relevant paper and dataset as: Stokanović, S., Sodnik, J., and Miljković, N., "From Raw Gaze to Meaningful Features: Assessment of Visual Behavior in Driving Simulator." ArXiv, 2025, https://doi.org/10.48550/arXiv.2511.02689. Stokanović, S., Sodnik, J., and Miljković, N., 2026. Software for Comprehensive Eye Tracking Feature Extraction [Software]. Zenodo, https://doi.org/10.5281/zenodo.18660155, GitHub: https://github.com/smiljastokanovic/Eye-Tracking-Parameters-Calculation, Accessed on: February 26th, 2026. Stokanović, S., Sodnik, J., and Miljković, N., 2026. Eye Tracking Dataset for Comprehensive Eye Tracking Feature Extraction [Dataset]. Zenodo, https://doi.org/10.5281/zenodo.18803236. # Eye-Tracking Parameters Dataset This repository contains **eye-tracking data** collected for gaze analysis. The dataset includes **2D and 3D gaze coordinates**, **eye origins**, **eye directions**, and **pupil sizes** for both the left and right eyes. --- ## Dataset Format The data is stored in **txt files**. Each row corresponds to a timestamped measurement. ### Columns | Column Name | Description | |------------------------|-------------| | `timestamp` | Time of recording (seconds) | | `gaze2d_x`, `gaze2d_y` | 2D gaze coordinates on the screen (normalized units) | | `gaze3d_x`, `gaze3d_y`, `gaze3d_z` | 3D gaze coordinates in space (scene coordinates) | | `eyeleft_origin_x`, `eyeleft_origin_y`, `eyeleft_origin_z` | 3D position of the left eye origin | | `eyeleft_dir_x`, `eyeleft_dir_y`, `eyeleft_dir_z` | 3D gaze direction vector of the left eye | | `pupil_left` | Pupil diameter for the left eye | | `eyeright_origin_x`, `eyeright_origin_y`, `eyeright_origin_z` | 3D position of the right eye origin | | `eyeright_dir_x`, `eyeright_dir_y`, `eyeright_dir_z` | 3D gaze direction vector of the right eye | | `pupil_right` | Pupil diameter for the right eye | --- ## Notes on Data - Some rows at the beginning may contain **`NaN` values** where the eye-tracker did not record valid data. - `NaN` values are common in initial samples or during blinks. - Users should handle missing data appropriately before analysis.本仓库收录了来自Tobii Pro Glasses 3眼动追踪仪（eye tracker）的眼动信号。本研究于搭载运动模拟平台的驾驶模拟器构建的仿真驾驶环境中开展。该数据集包含26名健康受试者的眼动信号。实验任务包含基线任务、正常路况驾驶任务（简称Ride）以及雾天驾驶任务（简称Fog）。本数据集的采样频率（sampling frequency）为100 Hz，采集分辨率（resolution）为1920×1080像素。 文件命名规范为：ID_gazedata_output。 为便于测试下载与预览，我们将所有眼动信号打包为.rar压缩包。 ## 数据集组成 - Eyetracking-database.rar：数据文件，文本格式 - README.txt：数据文件元数据，文本格式 若您的研究或教学工作中使用了本数据集，请一并引用如下论文与数据集资源： Stokanović, S., Sodnik, J., 与 Miljković, N., 《从原始视线数据到有效特征：驾驶模拟器中视觉行为的评估》，ArXiv, 2025, https://doi.org/10.48550/arXiv.2511.02689. Stokanović, S., Sodnik, J., 与 Miljković, N., 2026. 综合眼动追踪特征提取软件[软件]. Zenodo, https://doi.org/10.5281/zenodo.18660155, GitHub: https://github.com/smiljastokanovic/Eye-Tracking-Parameters-Calculation, 访问时间：2026年2月26日. Stokanović, S., Sodnik, J., 与 Miljković, N., 2026. 综合眼动追踪特征提取眼动数据集[数据集]. Zenodo, https://doi.org/10.5281/zenodo.18803236. # 眼动参数数据集 本仓库收录了用于视线分析的**眼动追踪数据（eye-tracking data）**。该数据集包含左右双眼的**二维与三维视线坐标（2D and 3D gaze coordinates）**、**眼球原点位置（eye origins）**、**视线方向向量（eye directions）**以及**瞳孔尺寸（pupil sizes）**。 --- ## 数据集格式 数据以**文本文件（txt files）**格式存储，每一行对应一条带时间戳（timestamp）的测量记录。 ### 字段说明 | 字段名 | 说明 | |-------------------------|----------------------------------------------------------------------| | `timestamp` | 记录时间（单位：秒） | | `gaze2d_x`、`gaze2d_y` | 屏幕二维视线坐标（归一化单位（normalized units）） | | `gaze3d_x`、`gaze3d_y`、`gaze3d_z` | 空间三维视线坐标（场景坐标系下的坐标（scene coordinates）） | | `eyeleft_origin_x`、`eyeleft_origin_y`、`eyeleft_origin_z` | 左眼眼球原点的三维位置 | | `eyeleft_dir_x`、`eyeleft_dir_y`、`eyeleft_dir_z` | 左眼视线方向的三维向量 | | `pupil_left` | 左眼瞳孔直径 | | `eyeright_origin_x`、`eyeright_origin_y`、`eyeright_origin_z` | 右眼眼球原点的三维位置 | | `eyeright_dir_x`、`eyeright_dir_y`、`eyeright_dir_z` | 右眼视线方向的三维向量 | | `pupil_right` | 右眼瞳孔直径 | --- ## 数据说明 - 数据集开头的部分行可能包含**NaN值**，对应眼动仪未采集到有效数据的时段。 - 初始采样片段或眨眼过程中常会出现NaN值。 - 用户需在分析前对缺失数据进行合理处理。