Eye Tracking¶

The instrumented measurement of eye position, eye motion, or point of gaze relative to a declared head, display, or scene frame, using calibration and qualified interpretation.

Version: v1 · 2026-09-28 · History
Domain-specific #: 9383
Domain group: Social Sciences
Origin domain: Psychology & Behavioral Sciences
Subdomains: Vision Science, Oculomotor Measurement, Experimental Psychology → Psychology & Behavioral Sciences
Aliases: Gaze tracking, Eye-gaze tracking, Oculography

Core Idea¶

Eye tracking turns signals from an eye into a time series of ocular position, movement, or estimated gaze. Eye-attached sensors, video-based optical systems, and electrooculography acquire different observables; video trackers commonly use pupil center and corneal reflections. A calibration model then maps sensor features into eye rotation or coordinates in a head, display, or world reference frame.

The record can be segmented into fixations, saccades, regressions, and scan paths for research or control. Those labels remain measurements or model-derived events, not direct readings of thought. Task instructions alter where people look, covert attention can differ from gaze, and the same fixation can support several cognitive interpretations. Valid use therefore joins spatial and temporal accuracy with an explicit inference boundary.

Structural Signature¶

Sig role-phrases:

Observed eye — Supplies the biological motion or orientation being measured. It is required target. Counterfactual: Head position or scene video without eye data is not eye tracking.
Sensing method — Acquires optical, attached-object, or electrical signals correlated with eye movement. It is required instrument. Counterfactual: Unaided impression lacks the instrumented measurement contract.
Reference frame — Defines gaze or rotation relative to head, display, or world scene. It is required coordinate context. Counterfactual: Coordinates are uninterpretable if their frame changes silently.
Calibration and mapping — Relates sensor features to eye orientation or point of gaze for a participant and setup. It is required measurement step. Counterfactual: Raw pupil pixels are not yet validated gaze coordinates.
Temporal samples and events — Records movement through time and may segment it into fixations, saccades, and related events. It is characteristic output. Counterfactual: A static eye image cannot supply a scan path.
Inference boundary — Separates measured ocular behavior from claims about attention, recognition, preference, or thought. It is required validity boundary. Counterfactual: Equating gaze with one mental state overclaims the measurement.

What It Is Not¶

Eye tracking is not head tracking alone; the eye's orientation or gaze must be measured.
It is not equivalent to visual-attention measurement because attention can move without the eyes and gaze can have ambiguous meaning.
A camera image of a face is not yet an eye-tracking record unless eye features are extracted and mapped to a stable frame.
Fixation duration does not by itself prove liking, recognition, confusion, deception, or any other unique mental state.
Closest near-miss. Attention tracking may use gaze as evidence, but covert attention and ambiguous fixation meanings prevent the two from being identical.

Scope of Application¶

Vision and reading research. Fixations, saccades, regressions, and task-dependent scan paths test accounts of visual behavior.
Human–computer interaction. Gaze records support usability studies and can act as an input channel.
Assistive control. Calibrated gaze can select targets or control interfaces when other motor channels are limited.
Applied observation. Design, driving, cartography, and other fields use gaze patterns with domain-specific validity controls.

Clarity¶

Reports should name device type, sampling rate, accuracy, precision, calibration, reference frame, event-detection rule, missing-data treatment, and participant task. A heat map without these details can hide whether a pattern comes from actual gaze, head motion, calibration drift, smoothing, or aggregation. Cognitive claims need independent theory or corroboration beyond the coordinate trace.

Manages Complexity¶

Eye tracking compresses high-frequency sensor signals into gaze coordinates and then into events or spatial summaries. Each layer removes detail and introduces assumptions: optical feature extraction, calibration geometry, fixation thresholds, and aggregation across people. Keeping those transformations visible makes a complex trace usable without treating a visualization as raw observation.

Abstract Reasoning¶

Choose a sensing method appropriate to the required precision, movement freedom, and participant burden.
Define the head, display, or world reference frame and perform participant-specific calibration.
Record signals with synchronized task and scene information while monitoring tracking loss and drift.
Map sensor observations to eye or gaze coordinates and quantify quality.
Derive fixations, saccades, or scan paths using a stated algorithm and thresholds.
Interpret patterns relative to task and corroborating evidence, keeping mental-state inference separate from measurement.

Knowledge Transfer¶

Eye tracking transfers among research, interface, assistive, and applied settings when ocular signals, calibration, coordinates, and quality controls remain intact. A cursor trail or head orientation can be correlated with gaze but is not literal eye tracking. The broader measurement pipeline transfers widely; the physiological target and gaze geometry make this domain-specific.

Cross-Domain Echoes¶

See how this entry connects to another domain.

A visible signal still needs an inference bridge

Examples¶

Canonical¶

A video tracker estimates pupil center and corneal reflection, calibrates those features to display coordinates, and records a participant's gaze samples while reading.

Mapped back: frame → display coordinates; mapping → participant calibration; output → gaze samples and fixations; sensor → infrared video; target → participant's eyes.

Applied / In Practice¶

A usability study compares scan paths under two tasks but avoids concluding that a long fixation uniquely means confusion, liking, or recognition.

Mapped back: context → assigned task; inference boundary → several cognitive explanations remain; measurement → fixation duration and location.

Structural Tensions¶

T1 — Measurement Sensitivity versus Intrusion And Ecological Validity. Eye-attached or tightly controlled systems can improve precision while altering behavior or limiting natural settings.

Diagnostic: What accuracy is gained, and what participant burden or scene constraint creates a new bias?

T2 — Rich Behavioral Trace versus Underdetermined Mental Meaning. A scan path is temporally detailed but does not uniquely identify cognition or covert attention.

Diagnostic: Which inference is directly measured, which is modeled, and what corroborating method supports it?

Structural–Framed Character¶

Eye Tracking is mixed. Instruments, geometry, calibration, sampling, and error are structural measurement features. Task choice, event thresholds, participant behavior, privacy expectations, and cognitive interpretation are framed by research and application practice. A technically precise trace can still support an invalid interpretation.

Structural Core vs. Domain Accent¶

The skeleton is instrumented mapping from a target attribute into a coordinate time series with uncertainty. Vision science supplies the eye, ocular features, rotations, gaze, fixations, saccades, and covert-attention boundary. Removing those components leaves general tracking or measurement.

This entry is a kind of Measurement.

Parent — Measurement (subsumption). Eye tracking necessarily maps ocular attributes through an instrument and calibration into frame-relative values with uncertainty.
Related — monitoring and attention. Repeated measurement supports monitoring, while attention is an inferential neighbor rather than an equivalent output.

Relationships to Other Abstractions¶

Current abstraction Eye Tracking Domain-specific

Parents (1) — more general patterns this builds on

Eye Tracking is a kind of Measurement Prime

Eye Tracking is a strict kind of Measurement: The instrumented measurement of eye position, eye motion, or point of gaze relative to a declared head, display, or scene frame, using calibration and qualified interpretation.

Hierarchy path (1) — routes to 1 parentless root

Eye Tracking → Measurement

Neighborhood in Abstraction Space¶

Eye Tracking sits in a crowded region of the domain-specific corpus (33^rd percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Visual Perception & Media Representation (20 abstractions)

Nearest neighbors

Posturography — 0.90
Digital Imaging — 0.89
Optical resolution — 0.89
Active perception — 0.88
Chartjunk — 0.88

Computed from structural-signature embeddings · 2026-10-08

Not to Be Confused With¶

Gaze estimation. Tell: Is the mapping operation often used inside eye tracking; eye tracking also includes acquisition, calibration, time series, and quality control.
Head tracking. Tell: Measures head pose and may assist gaze estimation but omits eye movement relative to the head.
Attention tracking. Tell: Makes a cognitive claim that gaze alone cannot uniquely establish.
Electrooculography. Tell: Is one sensing method for eye movement, not a synonym for the entire field or pipeline.

References¶

Frozen Wikipedia discovery revision: https://en.wikipedia.org/wiki/Eye_tracking (revision 1369021290).
Preserved source candidate: https://trove.nla.gov.au/work/12223957
Preserved source candidate: http://wexler.free.fr/library/files/yarbus%20%281967%29%20eye%20movements%20and%20vision.pdf
Preserved source candidate: http://www.learning-systems.ch/multimedia/forsch1e.htm
Preserved source candidate: http://www.learning-systems.ch/multimedia/eye%20movements%20problem%20solving.swf
Preserved source candidate: https://web.archive.org/web/20110706235011/http://www.learning-systems.ch/multimedia/eye%20movements%20problem%20solving.swf
Preserved source candidate: https://books.google.com/books?id=b_HSJKidixAC
Preserved source candidate: https://archive.org/details/mindseyecognitiv0000unse
Preserved source candidate: http://www.mmi-interaktiv.de/uploads/media/MMI-Interaktiv0303_SchiesslDudaThoelkeFischer.pdf

The frozen Wikipedia revision is discovery provenance. The retained source set was reviewed for identity, formal or operational relation, and scope. The encyclopedia's structural synthesis is bounded to those claims; a thin authority surface is recorded as a nonblocking source-strengthening repair rather than concealed.