Five universities recently announced new photographic technology:
FASTEST CAMERA EVER
With a laser that fires millions of pulses a second, the Serial Time-Encoded Amplified imaging technique captures images faster than any other camera – more than 6 million frames a second.
Developed at the University of California, Los Angeles, “Steam” is aimed at medical imaging, such as analyzing flowing blood samples. It eventually will record at more than 200,000 times faster than any standard video camera, the developers say.
SIMULTANEOUS ZOOM AND WIDE ANGLE
Researchers at Princeton claim to capture a wide-view image that maintains super-high resolution, taking a closer look at an object without narrowing the field of view; so, in effect, all parts of a scene will be “zoomed in” at the same time.
The technique passes light through a crystal nonlinear optical material, in which light rays mix in ways that transmit information that would otherwise be lost: Normal cameras record the color and brightness of light; this technique captures the property called phase, which measures the time and location of a wave peak. Phase is recorded in a hologram.
The initial image is distorted; but it is fixed with a computer algorithm, and then combined with data from a normal camera.
The process is a long way from SLRs, however: it’s being developed for microscopes, for better medical diagnostics.
SEEING “STAR TREK”
Carnegie Mellon University scientists have applied “state-of-the art algorithms in face detection, face tracking, and face recognition to 67 ‘Star Trek’ episodes.” We’re talking old school – the original TV series from the ’60s.
The process automatically extracts all visible faces, and clusters these into a small number of same-person groupings, the researchers say. It now recognizes frontal or near-frontal faces, but may soon work with nonfrontal shots as well.
Once the process found all the faces, a person took less than 5 minutes to assign names to all the main characters (Kirk, Spock, McCoy, Uhura, Scotty, Sulu, and Chekov) and a couple minor ones (Janice and Nurse Chapel) across all 67 episodes. They then graphed the relative face time for each of the main characters across all three seasons.
The technology is being commercialized by Carnegie Mellon spin-off Pittsburgh Pattern Recognition, which says the face mining technology “enables new video viewing experiences,” such as clickable navigation to scenes of interest, searching for specific people, hyperlinking scenes of interest, adding commentary synchronized with play points of interests, and summarizing occurrences of characters and people.
BRIGHT COLOR, LOW POWER DISPLAYS
Electrically switching the appearance of pigments may yield full-color “electronic paper” that, compared to other solutions, is “ahead by a wide margin in critical categories, such as brightness, color saturation, and video speed,” Nature Photonics reports.
Electrofluidic display technology developed by researchers at the University of Cincinnati uses transposition of pigment dispersions. The reflective display can be less than 15 microns thick – for flexible, rollable screens.
A new company, Gamma Dynamics, will commercialize the technology.
MAPPING 35 MILLION FLICKR PHOTOS
Scientists at the Cornell University Center for Advanced Computing analyzed the geotags on nearly 35 million photos taken by more than 300,000 photographers and uploaded to the Flickr photo sharing site.
The researchers found the most-photographed landmark in each of the top 20 most-photographed cities, according to photo density. The most-photographed cities were New York City, London, San Francisco, Paris, Los Angeles, and Chicago.
Most-photographed landmarks included the Eiffel Tower, Trafalgar Square, Big Ben, and Notre Dame. Perhaps it says something about Flickr users that the 28th most-photographed place in the world was the Apple Store in midtown Manhattan.
May 4, 2009
