December 13, 2006 – We all have improperly focused photos that we wish we could fix. Instead of merely hoping to turn back time, photographers can look to Stanford University research to help readjust focusing post-capture. Recent Stanford PhD graduate and project researcher Ren Ng will present, today, a plenoptic camera, also called the 4D light field camera or Digital Lens, at a Camera Owners of the Bay Area (COBA) meeting in San Francisco. The prototype plenoptic camera and software can change depth of field on a digital image after exposure.
Announced in October 2005, the plenoptic camera prototype quickly grabbed the attention of photographers, technophiles, and industry insiders. Since then, the 4D light field technology has been getting increasingly closer to becoming commercialized for everyday use.
The plenoptic camera is based on the work done by Massachusetts Institute of Technology (MIT) researchers John Wang and Edward Adelson in 1992. The new prototype consists of a 16-megapixel Contax 645 camera with added multiple microlenses in between the photo sensor and the main lens, thus turning a standard camera into a light field camera. The added microlens array measures light distribution on each ray of light, which is information that is normally lost on traditional cameras. By collecting that vital information through custom post-production software, the Stanford researchers were able to compute sharpness at different focal lengths.
Researchers could then extract a different pixel from the microlens image to change the point of view of the observer across the aperture plane, that is, altering the focal length of the “synthetic image,” according to a 2005 abstract “Light Field Photography with a Hand-Held Plenoptic Camera” by Ren Ng, Mathieu Bredif, Gene Duval, Marc Levoy, Mark Horowitz, and Pat Hanrahan. The new plenoptic camera can essentially turn a 2D photo into 4D.
What the research could do varies from everyday photography to larger scientific applications. The portable 4D camera may be particularly useful for photographers, giving them more flexibility before and after capture. A photographer could extend depth of field without reducing aperture. For example, if a user shot at the wide aperture f/4 to keep exposure time short and keep noise low, the user is usually left with an image with shallow depth of field, that is, a partially focused photo. With the new technology, the photographer could shoot at f/4 but still have an image that looks like it was shot at f/22 with all parts in focus.
“The result is a way of decoupling the traditional trade-off in photography between aperture size and depth of field,” stated the abstract. Photographers may no longer need to be restricted by the delicate balance between the two.
Another advantage of the digital refocusing technology is that it could improve image quality rather than simply increasing megapixel count. “Conventional cameras have already reached the limits of useful resolution for many common applications,” stated the abstract. Since resolution in this case is relative to the size of the microlens and not megapixel count, the plenoptic camera could finally end the manufacturer megapixel war with a new innovation.
The 4D camera and software could also function as a money-saver. Photographers could buy less expensive, slower lenses and then correct poorly focused images after the fact.
.jpg)
.jpg)
.jpg)
Online sharers could enjoy interactive photo albums. Sports photographers and security cameras could produce images shot in low light with more focal lengths. Medical researchers and scientists could extract more information from live subjects where depth of field is particularly shallow, stated the abstract.
Best of all, digital refocusing technology only requires a “simple optical modification to existing digital cameras,” according to the abstract. “To the photographer, the plenoptic camera operates exactly like an ordinary hand-held camera." The 4D light field camera needs only internal alterations to the 2D photo sensor with no changes necessary to exterior bodies, which is a benefit for photographers who are accustomed to the design of their cameras.
The Stanford researchers tested the 4D light field technology on hundreds of light field photos including portraiture, action, and macro close-up shots. The final resolution of the images was 300 x 300 pixels, which is equal to the microlens array, although Ren Ng notes on a Stanford website that the final output images were with the prototype only. The resolution could be changed relative to the microlens.
Light field photography is already in the process of being commercialized by Refocus Imaging, Inc., a company headed by Ren Ng, which will market the new technology. More information on Refocus Imaging can be found at http://www.refocusimaging.com/.
