Brian's Brain

Continued from 'Imaging Beyond Pixels: Future Forecasts'....

What other breakthroughs caught my eye as I was researching this article? In a recent blog post, I described the mathematical theory behind the Healing Brush features in latest generations of Adobe's Photoshop. An article in the October 2006 issue of Scientific American Magazine (you can probably tell by now that I've been planning this article for a while!), entitled 'Impaint By Numbers', expands the image repair concept to encompass video. Although the concept is presented as a post-capture repair process, there's nothing that inherently prevents its use during video capture as well.

Optics innovations were a key focus of my print piece. The February issue of PC World Magazine showcases yet another one, this time from Adobe. In a short writeup entitled 'Megapixels On the Rise', Alan Stafford writes, "Adobe's Advanced Technology Labs showed us a prototype lens that incorporates 20 smaller lens elements (each with 10-megapixel resolution, for a total of 200 megapixels) in a honeycomb pattern. Each element captures the same image but with a different depth of field and focus; you can then select the image with the depth and focus you want, or use Adobe's skunk-works software to combine portions of each element."

Taking my article's 'today's cameras' megapixel counts are overkill for most applications' stance to the extreme, researchers at Rice University have developed a prototype single-pixel sensor camera that iteratively captures portions of a scene (followup Slashdot coverage here). It employs a DMD (Digital Micromirror Device) from Texas Instruments, conceptually similar to what's inside a DLP (Digital Light Processing) display or projector, to focus segments of the scene on the sensor. Rice's Dharmpal Takhar presented at the TI Developer Conference last week; this slide (JPG) from his presentation explains the concept while this one shows the setup. Practical applications for the technology are, frankly, unclear to me; like the high resolution camera back I saw at PMA, the single-pixel camera has extremely long exposure times.

Finally, applying singularity to image storage, we turn to the University of Rochester, whose researchers have figured out how to archive (and later retrieve) an entire (low-resolution, but still....) image within a single photon (commentary from Slashdot). Relying on the fact that light exhibits both particle and wave behaviours, the group figured out how to pass a single photon, carrying the 'shadow' of the entire image, through a stencil; they then used a tube of cesium gas to slow its speed. Quoting from the writeup, " Howell has so far been able to delay light pulses 100 nanoseconds and compress them to 1 percent of their original length. He is now working toward delaying dozens of pulses for as long as several milliseconds, and as many as 10,000 pulses for up to a nanosecond."