A grad student in computer science at Stanford University has adapted the 16MP Kodak CCD sensor in a Megavision FB4040 digital back to work as a "light field camera," thereby enabling the point of focus in a picture to be adjusted in software after the picture is taken.
A specially-designed 90,000 microlens array that alters the way the sensor captures light, and custom software to process the resulting digital files, together provide the option of either expanding depth of field after the fact - making, for example, a photo shot at f/4 appear to have the depth of field of one taken at f/22 - or simply changing the point of focus to a different distance (within limits).
Ren Ng , the Stanford PhD student who led up the project, mated the modified Megavision back to a Contax 645 and used 140mm f/2.8 and 80mm f/2.0 lenses in developing the technology. A Stanford University press release says we may eventually see this type of microlens array (which is radically different in its design than the microlens arrays common to image sensors today) inside cameras aimed at the security and surveillance market.
Because each lenslet in the array is considerably larger than each 9µm Kodak sensor pixel, the real-world resolution of the Stanford light field camera is only a fraction of what one would normally expect from a 16MP capture device. This will likely limit the technology's usefulness in photography segments that require medium-to-high-resolution digital files, at least until the technology evolves considerably.
Thanks to Ken Geiger for bringing this to our attention.