ETH-XGaze Extreme-Pose Gaze Estimation Dataset
High-resolution multi-camera face images with calibrated gaze targets, extreme head poses, and varied illumination.
At a glance
| Field | Value |
|---|---|
| Short name | eth_xgaze |
| Full name | ETH-XGaze Extreme-Pose Gaze Estimation Dataset |
| Primary category | eye_tracking |
| Contained modalities | eye_tracking |
| Tasks | gaze_estimation, regression |
| Samples | 1,083,492 |
| Classes | Not reported (Not reported) |
| Splits | all |
| Size | 130.0 GB |
| Source-stated terms | CC BY-NC-SA 4.0 with additional dataset conditions |
| Normalized terms | cc-by-nc-sa |
| Descriptive screening label | Explicit noncommercial clause recorded; check source |
| Terms scope | dataset_files |
| Access friction | controlled_or_manual |
| Route backend | Manual (upstream-gated) |
| Availability | available (checked 2026-07-21) |
| Acquisition support | manual_access_blocked |
| Legacy sample-loader status | Metadata and access only |
Notes
Contains 1,083,492 images from 110 participants. The official page offers approximately 130 GB 224-pixel face patches, 497 GB 448-pixel patches, or about 7 TB of raw images on request. The linked completed license adds conditions to CC BY-NC-SA 4.0.
Access preflight and acquisition
- CLI
- Python
# This route requires upstream human action; no transfer starts.
eyehub download eth_xgaze --data-dir ./data --dry-run --json
# Follow the official instructions shown by preflight.
from eyedatahub.acquisition import preflight_dataset
from eyedatahub.datasets.registry import REGISTRY
ds = REGISTRY.get_dataset('eth_xgaze')
print(preflight_dataset(ds, './data')) # returns manual_access_blocked
Upstream page: ait.ethz.ch/xgaze
Source-term evidence: ait.ethz.ch/xgaze
Loader status
This catalog record provides metadata and access instructions, but it does not yet include a standard DatasetSample loader. Inspect the source file structure or contribute a loader before using it in a training pipeline.
Citation
- BibTeX
- Plain text
@misc{eth_xgaze,
title = { ETH-XGaze Extreme-Pose Gaze Estimation Dataset },
note = { Zhang X, Park S, Beeler T, Bradley D, Tang S, Hilliges O. ETH-XGaze: A Large Scale Dataset for Gaze Estimation under Extreme Head Pose and Gaze Variation. ECCV. 2020 },
year = { 2020 },
url = { https://ait.ethz.ch/xgaze },
}
Zhang X, Park S, Beeler T, Bradley D, Tang S, Hilliges O. ETH-XGaze: A Large Scale Dataset for Gaze Estimation under Extreme Head Pose and Gaze Variation. ECCV. 2020.
Source-stated terms
- Raw source string: CC BY-NC-SA 4.0 with additional dataset conditions
- Normalized category:
cc-by-nc-sa - Apparent scope:
dataset_files - Descriptive screening label: Explicit noncommercial clause recorded; check source
⚠️ Source-stated terms, scope, and normalized labels are curation metadata, not legal advice or a permission finding. Review the current official source before transfer or reuse.
Related datasets with shared modalities
- teyed: TEyeD Real-World Eye-Tracking Dataset (20,666,096 records,
unknown) - gaze_capture: GazeCapture Mobile Eye-Tracking Dataset (2,445,504 records,
unknown) - mpii_gaze: MPIIGaze Appearance-Based Gaze Estimation Dataset (213,659 records,
cc-by-nc-sa) - gaze360: Gaze360 Physically Unconstrained Gaze Dataset (172,000 records,
unknown) - lpw: LPW Labelled Pupils in the Wild (130,856 records,
research-only) - glaucoma_eye_movements: Eye Movements of Glaucoma Patients with Asymmetrical Visual Field Loss (270 records,
cc-by)