Skip to main content

GazeCapture Mobile Eye-Tracking Dataset

Crowdsourced iPhone and iPad face videos with screen-fixation coordinates for appearance-based mobile gaze estimation.

At a glance

FieldValue
Short namegaze_capture
Full nameGazeCapture Mobile Eye-Tracking Dataset
First published2017-03
Publication date precisionmonth
Publication date evidencegithub.com/CSAILVision
Publication date source fieldOfficial GitHub README: History
Publication date reviewed2026-09-11
Primary categoryeye_tracking
Resource rolecurrent_dataset
Dataset familygaze_capture
Contained modalitieseye_tracking
Tasksgaze_estimation, regression
Primary reported quantity2,445,504 frames
ClassesNot reported (Not reported)
Splitsall
SizeNot reported
Source-stated termsUnknown; registration required and source terms must be checked
Normalized termsunknown
Descriptive screening labelUnknown or unclear; do not assume permission
Terms scopeunknown
Access frictionself_service_authenticated
Route backendManual (upstream-gated)
Availabilityavailable (checked 2026-07-21)
Acquisition supportguided_instructions_only
Legacy sample-loader statusMetadata and access only

Reported quantities

RoleCountUnitScopeBasisEvidence
Primary2,445,504framesLabeled mobile-camera framesofficial_source_descriptiongazecapture.csail.mit.edu
Additional1,474participantsCrowdsourced participantsofficial_source_descriptiongazecapture.csail.mit.edu

Counts retain their source-reported units. Additional rows can describe components, paired items, or derivative copies and are not automatically added to the primary quantity.

Notes

The primary paper reports 2,445,504 frames with fixation locations from 1,474 participants. The official site requires account registration; verify the current agreement before reuse.

Access information and download

# Read-only preflight
eyehub download gaze_capture --data-dir ./data --dry-run --json

# Download, only when preflight reports supported behavior
eyehub download gaze_capture --data-dir ./data

Upstream page: gazecapture.csail.mit.edu/dataset.php

Source-term evidence: gazecapture.csail.mit.edu/dataset.php

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

@misc{gaze_capture,
title = { GazeCapture Mobile Eye-Tracking Dataset },
note = { Krafka K, Khosla A, Kellnhofer P, et al. Eye Tracking for Everyone. CVPR. 2016:2176-2184. arXiv:1606.05814 },
year = { 2016 },
url = { https://gazecapture.csail.mit.edu/dataset.php },
}

Source-stated terms

  • Raw source string: Unknown; registration required and source terms must be checked
  • Normalized category: unknown
  • Apparent scope: unknown
  • Descriptive screening label: Unknown or unclear; do not assume permission

⚠️ 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.

Similar resources by shared modality

  • teyed: TEyeD Real-World Eye-Tracking Dataset (20,666,096 images, unknown)
  • eth_xgaze: ETH-XGaze Extreme-Pose Gaze Estimation Dataset (1,083,492 images, cc-by-nc-sa)
  • mpii_gaze: MPIIGaze Appearance-Based Gaze Estimation Dataset (213,659 images, cc-by-nc-sa)
  • gaze360: Gaze360 Physically Unconstrained Gaze Dataset (172,000 frames, unknown)
  • lpw: LPW Labelled Pupils in the Wild (130,856 frames, research-only)
  • glaucoma_eye_movements: Eye Movements of Glaucoma Patients with Asymmetrical Visual Field Loss (270 participants, cc-by)
  • mendeley_sub_cone_visual_resolution_by_active: Data from: Sub-cone visual resolution by active, adaptive sampling in the human foveola (16 participants, cc-by)
  • asd_eye_tracking: Autism Spectrum Disorder Eye-Tracking Dataset (Not reported, cc-by)