Exemplary applications of our camera add-on: Top Row: Multispectral imaging: unprocessed output (left), a spectral stack of images after processing (middle), a neutral image relit with a flat spectrum (right). Middle Row: High Dynamic Range Imaging: unprocessed output (left), a simulated exposure sweep of contrast ratio 100 : 1 (middle), a tone-mapped version of the HDR image (right). Bottom Row: Light-field imaging: unprocessed output (left), virtual refocusing on foreground (middle) and background (right).

Abstract

We propose a non-permanent add-on that enables plenoptic imaging with standard cameras which we refer to as KaleidoCamera. Our design is based on a physical copying mechanism that multiplies a sensor image into a number of identical copies that still carry the plenoptic information of interest. Via different optical filters, we can then recover the desired information. A minor modification of the design also allows for aperture subsampling and, hence, light-field imaging. As the filters in our design are exchangeable, a reconfiguration for different imaging purposes is possible. We show in a prototype setup that high dynamic range, multispectral, polarization, and light-field imaging can be achieved with our design.

Citation

A. Manakov, J. Restrepo, O.Klehm, R. Hegedus, E. Eisemann, H.-P. Seidel, I. Ihrke. 2013. A Reconfigurable Add-on for Multi-Spectral, High-Speed, Polarization and Light-Field Imaging. ACM Trans. Graph. 32, 4 (Proc. SIGGRAPH), 2013.

Bibtex

@article{KaleidoCamera-Manakov2013,
  author = {Manakov, Alkhazur and Restrepo, John F. and Klehm, Oliver and Heged\"{u}s, Ramon and Eisemann, Elmar and Seidel, Hans-Peter and Ihrke, Ivo},
  title = {A Reconfigurable Camera Add-on for High Dynamic Range, Multi-Spectral, Polarization, and Light-Field Imaging},
  journal = {ACM Trans. Graph. (Proc. SIGGRAPH 2013)},
  issue_date = {July 2013},
  volume = {32},
  number = {4},
  month = jul,
  year = {2013},
  issn = {0730-0301},
  pages = {47:1--47:14},
  articleno = {47},
  numpages = {14},
  url = {http://doi.acm.org/10.1145/2461912.2461937},
  doi = {10.1145/2461912.2461937},
  acmid = {2461937},
  publisher = {ACM},
  address = {New York, NY, USA},
  keywords = {computational optics, computational photography},
}
[KaleidoCamera-Manakov2013.bib]

Paper and Video

Paper [KaleidoCamera-Manakov2013.pdf]
Video [KaleidoCamera-Manakov2013.mp4]

Presentation

Here you can download the presentation for the talk given at SIGGRAPH 2013 and the video recording of the talk.

Presentation materials [Presentation]
Video of the talk [SIGGRAPH talk]

Press

Results

We show results for the light field design, as well as the filter-based design with filters for HDR, multispectral, and polarization photography.

Light field

[show Light field results]

Captured image

In the light-field configuration of the KaleidoCamera the diffuser is removed, which enables sub-sampling of viewing directions. The central sub-image is a direct view to the scene (without reflections). The other sub-images are the other views of the scene, flipped because of the mirror-reflection. The corner sub-images are flipped twice because of a double mirror-reflection. Please pay attention to the parallax, for instance, look at the left of the pirate. In the central sub-image it is slightly occluded by Buddha's right shoulder, in the left image in the second row the leg is fully visible, and in the right image it is ocluded.

Refocusing Sequence

If you seen an error message, please try reloading the page.

High Dynamic Range

[show High Dynamic Range results]

Captured image

In the filter-based configuration of the KaleidoCamera works with the diffuser and interchangeable filters, in this case neutral density filters. The diffuser functions as a back projection screen and enables physical image copying avoiding parallax.

Simulated exposure sweep

Multispectral

[show Multispectral results]

Captured image

Under filter-based configuration with wide-band spectral filters we are able to reconstruct the spectrum of the scene.

Simulated color sweep

In our experiment we have used two artificial roses and a natural one. The artificial roses' spectrum differs from the spectrum of the natural rose in the blue region as we can see from the second image in the top row.

Polarization

[show Polarization results]

Captured image

For polarization imaging it is essential to use a polarization-preserving diffuser, such type of diffusers are commonly used for polarization-based stereo projection. In this case we use linear polarization filters with orientations of 0, 45, 90 degrees. We are able to capture Stokes parameters for every pixel. The logo on the left (central sub-image) is shown on the screen, whereas the logo on the right is printed on paper. The paper acts as a diffuser and destroys the LCD’s light polarization. Thus the logo displayed on the screen is polarized, and the one on the paper is depolarized.

Virtually rotated polarizer

We simulate the image perceived by the eye looking through linear polarizer (in different rotation positions) at the scene. In the position it is possible to cancel polarized light out by the filter (left-most image in the top row).

Datasets

Here we provide one dataset for each plenoptic function dimension we are able to capture with the KaleidoCamera. Full datasets are available on request however are not being made directly downloadable due to hosting constraints.

Overview Document [README]
Code [Code for loading .ppm images]
[Code for loading .depthmap images] (only for light field data necessary)
Light-field [Light-field]
High Dynamic Range [HDR]
Multispectral [Multispectral]