Questions and Concepts

Epsilon to Coded Photography

Computational photography extends traditional cameras through computation, smart lights, and novel optics.

Epsilon photography synthesizes omnipictures via the multiple capture, single image (MCSI) paradigm — scene properties are encoded across a few photographs with slightly varied camera parameters.

Coded photography encodes scene information through modulated optics (aperture, shutter, illumination), then computationally decodes richer representations. Key foundations:

• Fourier transform and Fourier optics

• Optical heterodyning

• Helmholtz reciprocity

Light Field

The light field describes the amount of light traveling in every direction through every point in space.

The plenoptic illumination function expresses the image of a scene from any viewing position, angle, and time. Since rays are parameterized by three spatial coordinates \((x, y, z)\) and two angles \((\theta, \phi)\), it is a five-dimensional function.

Computational Illumination

Computational illumination asks: what spatio-temporal modulations of light best reveal the visually important contents of a scene?

• Changing color filters, tunable wavelength filters, or diffraction gratings

• Better lighting control during capture enables richer scene representations

• Reference

Epsilon photography approach: Each camera setting records partial scene information; the final image is reconstructed by computationally merging multiple observations, combining the best features from each.