← Back to catalogue
Published

light field

vr.tr.light-field · thing-q593161

Enable an agent to recognise a light field, assess its directional and spatial coverage, and decide which viewing, refocusing or measurement operations its evidence supports.

Thing Registry Cross-cutting context XCT.QLT

Bundle → Layer → Finding → Questions Filled

6 bundles · 11 layers · 16 findings · 27 questions

Definition and instance Establish which meaning of light field applies and what constitutes an instance.

Radiance convention

Identify the quantity and independent variables used by the adopted definition.

Position-direction definition

Record the adopted radiance definition and its authority. Levoy and Hanrahan define light fields through position and direction: [Light Field Rendering](https://www.graphics.stanford.edu/papers/light/).

  1. Which definition and source govern this instance, and what quantity does the function return? definition
  2. Are wavelength and time explicit coordinates, fixed conditions or integrated variables? boundary

Instance boundaries

Separate the field being described from its captures and computational encodings.

Field versus representation

Record whether an instance denotes a physical radiance distribution, a sampled observation or an inferred representation, and how these are related.

  1. Does this instance denote the optical distribution, a particular capture or a reconstruction? definition
  2. Which changes in scene, illumination or time create a different field instance? boundary
Ray domain and transport Specify ray identities, coverage and the physical assumptions behind their use.

Ray coordinate system

Describe how coordinates identify oriented rays.

Coordinate interpretation

Record reference frames, orientation, units and any reference planes needed to interpret ray coordinates.

  1. How does a sample address map to a physical ray, including its direction of travel? definition
  2. Which rays are excluded or singular under the chosen parameterisation? boundary

Transport assumptions

State where a reduced ray representation is valid.

Four-dimensional reduction

Record the region supporting a four-dimensional ray representation and the assumptions permitting position along a ray to be omitted. The classic formulation uses free-space regions: [Light Field Rendering](https://graphics.stanford.edu/projects/lightfield/gamma-corrected/).

  1. What evidence supports treating radiance as invariant along the represented ray segments? boundary
  2. Where do occlusion, attenuation, emission, scattering or refraction invalidate the adopted transport simplification? boundary
Radiometric, spectral and temporal meaning Make light-field values comparable and physically interpretable.

Radiometric values

Track units, response corrections and unusable measurements.

Radiance or encoded values

Require an explicit distinction between physical radiance and relative or display-encoded channel values.

  1. What units, exposure normalisation and response transformation define each value? measurement
  2. Which samples are clipped, below the noise floor or affected by unresolved gain differences? measurement

Spectral and temporal compatibility

Describe the wavelength and time intervals contributing to samples.

Compatible ray observations

Record whether observations can be treated as samples of one field under the intended use.

  1. Which spectral responses and exposure intervals contribute to each group of ray samples? measurement
  2. What evidence supports combining sequential views despite possible scene motion or illumination change? provenance
Sampling and reconstruction evidence Describe observed ray support and the evidence behind reconstructed values.

Ray sampling support

Record sampling density, footprints and coverage gaps.

Measured ray coverage

Describe the actual spatial-angular sample pattern, including finite integration footprints and missing observations.

  1. What are the spatial and angular sample spacings and footprints across the supported domain? measurement
  2. Which requested rays fall between measured samples or outside observed coverage? boundary

Calibration and inference

Connect ray assignments and reconstructed values to their supporting evidence.

Sample lineage and error

Record geometric calibration, value provenance and validation for measured, simulated, interpolated or learned samples.

  1. Which calibration records establish ray locations and directions, and with what residual error? provenance
  2. Which values are inferred, and what held-out observations or other evidence bound their error? measurement
Supported operations Determine which transformations and inferences the particular field can support.

View and aperture queries

Assess requested viewpoints and computational focus settings against available rays.

Rendering query admissibility

Record the ray support and validation required for a requested viewpoint, focus setting and synthetic aperture.

  1. Do the requested viewpoint and aperture require rays outside the calibrated support? action
  2. Which focus and aperture settings meet the task's stated sharpness and artifact tolerances? measurement

Geometric inference

Assess what directional variation permits an agent to infer about scene structure.

Disparity and depth conditions

Require geometric inference to state its calibration, correspondence and appearance assumptions and identify ambiguous regions.

  1. Which calibration and correspondence assumptions support converting angular variation into disparity or metric depth? definition
  2. Where do occlusion, weak texture, reflections or transparency make the inferred geometry ambiguous? boundary
  3. Should the agent return an uncertainty interval, request additional views or withhold a depth estimate? action
Evidence and external alignment What the world already says about this thing, gathered so the model can be checked against it.

Reported evidence

Findings from the breadth pass, kept separate from the structural claims.

Check these first

Recalled without web access and unsourced; every item is a lead to verify.

  1. Which of these check these first hold for the sense of light field this model covers, and on what evidence? provenance

Kinds and varieties

Recalled without web access and unsourced; every item is a lead to verify.

  1. Which of these kinds and varieties hold for the sense of light field this model covers, and on what evidence? provenance

Real-world use

Recalled without web access and unsourced; every item is a lead to verify.

  1. Which of these real-world use hold for the sense of light field this model covers, and on what evidence? provenance

Typical measurements

Recalled without web access and unsourced; every item is a lead to verify.

  1. Which of these typical measurements hold for the sense of light field this model covers, and on what evidence? provenance

Failure modes and hazards

Recalled without web access and unsourced; every item is a lead to verify.

  1. Which of these failure modes and hazards hold for the sense of light field this model covers, and on what evidence? provenance

Neighbouring kinds and how to tell them apart

Recalled without web access and unsourced; every item is a lead to verify.

  1. Which of these neighbouring kinds and how to tell them apart hold for the sense of light field this model covers, and on what evidence? provenance

Classifiers Filled

Family
Thing Registry
Category
Cross-cutting context
Entry kind
thing
Plane
XCT
Domain
XCT.QLT

What it is Filled

A light field describes the radiance of light as a function of position and propagation direction, usually at a fixed time and wavelength or over a specified spectral band.

In scope

  • Definitions distinguishing a general position-and-direction radiance function from a reduced four-dimensional ray representation
  • Ray coordinates, spatial boundaries, directional coverage and propagation assumptions
  • Spectral, temporal and radiometric interpretation of field values
  • Measured, simulated and inferred light-field samples and their calibration
  • Conditions supporting view synthesis, computational refocusing and geometric inference

Out of scope

  • Construction and operation of plenoptic cameras, camera arrays and light-field displays
  • Full electromagnetic fields, optical phase and wave-interference models
  • Scene geometry, material reflectance and illumination sources as independently modelled entities
  • Neural-network architectures and volumetric scene representations used to generate ray observations
  • Image-file formats, storage infrastructure and compression implementations

Why it exists Filled

Enable an agent to recognise a light field, assess its directional and spatial coverage, and decide which viewing, refocusing or measurement operations its evidence supports.

Distinguishing features Filled

  • A light-field representation associates values with both ray position and ray direction; a conventional photograph alone does not establish independently sampled variation across both.
  • An environment map records directional appearance from a reference location or under an explicit distant-scene approximation; it does not generally establish spatially varying ray values.
  • A depth map records geometric distance rather than directional radiance, even when both are derived from the same capture.
  • A radiance light field does not by itself record optical phase or the electric and magnetic field components needed for wave-optics analysis.
  • A model must declare whether its values describe radiance along rays or local emission and attenuation that require a transport calculation to produce ray observations.

What robots and AI may and may not do Filled

Must not

  • Capture light field images of people without their consent where consent is required.
  • Use refocusing or depth recovery to reveal details people meant to keep out of view.
  • Present synthetic views from light field data as unaltered photographs.

Only with a human decision

  • Using light field capture in surveillance or identification of people.

May

  • Explain the light field concept and its use in imaging, rendering and displays.
  • Capture or process light field data the agent is authorized to handle.

Moral aspects Filled

  • Light field capture records depth and multiple views, making images more revealing than ordinary photographs.
  • Synthetic views can be mistaken for real scenes.

Who is affected

  • People captured in images
  • Viewers
  • Researchers

Owners Filled

Steward

Nobody owns it; it is a concept of optics.

Links to other meta-models Filled

parent

  • Q11348 - registry parent class

What else AI and robots need to interact with it Filled

Identity and identifiers required Filled

  • Vercy registry: vr.tr.light-field
  • Wikidata: Q593161 (https://www.wikidata.org/wiki/Q593161)

Direct properties not applicable Not applicable

  • Definition and retained dimensions: Position-direction function; four-dimensional ray function; declared spectral or temporal extension
  • Ray parameterisation: Position plus direction; two-plane coordinates; another explicitly defined ray chart
  • Spatial and angular support: Spatial extent in m; directions in rad or unit vectors; supported solid angle in sr
  • Radiometric interpretation: Band-integrated radiance in W m^-2 sr^-1; spectral radiance with declared wavelength unit; calibrated channel values; relative encoded values
  • Sampling footprint: Spatial and angular spacing and integration footprints, with units defined by the ray chart
  • Temporal and spectral support: Capture times and exposure intervals in s; wavelength bands in nm; channel response functions
  • Propagation regime: Declared homogeneous transparent region; participating medium; refractive region; unresolved
  • Sample evidence origin: Links from ray samples or regions to measurements, simulation runs, reconstruction methods and calibration records
  • Operational validity: Supported, conditionally supported, unsupported or unassessed for each requested operation

Plane XCT: no invented physical properties.

Recognition not applicable Not applicable

Not applicable

A light field is a mathematical description whose identity cannot be established from visual appearance alone.

Capabilities and actions required Filled

  • Query directional radiance or channel values at supported ray coordinates.
  • Reparameterise rays while recording coverage loss and resampling effects.
  • Synthesize viewpoints with explicit limits on interpolation and extrapolation.
  • Compute refocused images or aperture integrations where ray coverage permits.
  • Estimate disparity or depth with stated correspondence assumptions and uncertainty.
  • Identify missing ray coverage and specify additional capture requirements.

Hazards and failure modes required Filled

  • Sampling errors cause artefacts that can be mistaken for real details.
  • Large datasets raise storage and privacy risks.

Standards and interfaces required Filled

  • The International Commission on Illumination publishes CIE S 017/E:2020, the International Lighting Vocabulary, which standardizes radiometric terminology including radiance.

Context of use required Filled

  • Light fields are used in computational photography, optical measurement and computer graphics to describe how radiance varies with position and direction.
  • Sampled light fields can support changes in viewpoint or computational refocusing when their coverage and resolution are sufficient.

Sources Filled

  1. Wikidata item Q593161: light field - identity and sense of the item
  2. Wikipedia: Light field - general description of the item

Open questions

  • Does vr.tr.light-field intend the optical radiance concept, the four-dimensional computational-photography convention or a broader electromagnetic meaning?
  • Which authoritative definition should govern the relationship between light field and plenoptic function in this registry?
  • Should time, wavelength and polarisation be optional extensions within this entry or relations to neighbouring models?
  • How should fields in participating or refractive media be represented without implying the free-space four-dimensional reduction applies?
  • Which task-specific validation criteria should determine acceptable view synthesis, refocusing and geometric inference?

Machine files

Provenance

thing registry research (pass 2) · unreviewed

Built from: models/things/publications/thing-q593161/spec.json