Light-field photography captures not just the intensity of light at each sensor point but also the direction of incoming rays. This additional directional information, known as the light field, allows computational operations after capture that traditional cameras cannot perform. Ren Ng, Stanford University, led the move of this concept into consumer devices, and Marc Levoy, Stanford University, developed many of the underlying rendering and refocusing algorithms that make practical use of recorded light fields.
How light fields are captured
A common implementation places a microlens array between the main lens and the sensor. Each microlens images the scene from a slightly different angle so that individual sensor pixels sample different ray directions. Recording that angular variation effectively samples the plenoptic function across spatial and directional dimensions. The raw data therefore contains a four dimensional representation of light across two spatial and two angular coordinates. This hardware arrangement trades some spatial detail for directional data, producing perceived lower native resolution in exchange for post-capture control.
Computational refocusing
Refocusing takes advantage of the recorded angular information. A refocusing algorithm synthetically shifts and integrates rays that correspond to a chosen focal plane. Conceptually the algorithm reprojects rays so they converge at a new depth and then sums their contributions to form an image that is sharp at that depth. Marc Levoy, Stanford University, described practical implementations where shifting and summing subimages created user-selectable focal depths and allowed estimation of a depth map from parallax across microlens subimages. The same directional data enables synthetic aperture effects where blending rays from different directions increases background blur or enhances depth of field.
Trade-offs and consequences
The primary consequence for practitioners is a change in photographic workflow. Photographers gain post-capture flexibility to adjust focus and depth effects, reducing the need for repeat takes in dynamic or inaccessible settings. Scientific and cultural heritage imaging benefit because refocusing and depth recovery can reveal features without retouching fragile artifacts. Environmental monitoring and microscopy adopt light-field capture for three dimensional measurement where re-imaging is impractical. The trade-offs include reduced native spatial resolution and increased computational load, which affect storage and processing demands. There are also cultural and ethical considerations when the same post-capture capabilities are applied to surveillance or forensics, since the ability to refocus after the fact alters expectations about when and how photographic evidence is acquired.