Aug 9, 2011 09:08 GMT  ·  By
Not only do the images produced by GelSight have a resolution that previously required expensive laboratory equipment, but they're 3D, too
   Not only do the images produced by GelSight have a resolution that previously required expensive laboratory equipment, but they're 3D, too

A collaboration of researchers from the Massachusetts Institute of Technology (MIT) Department of Brain and Cognitive Sciences announces the development of a simple, yet advanced portable imaging system, which is capable of achieving resolution previously thought reserved to larger equipment.

Until the new study was published, scientists did not think that it was possible to obtain such a high level of magnification and resolution with an instrument that is not large or expensive, and does not require a scientific lab to operate.

Developing this device was made possible by an intricate combination of clever physical interfaces with computer-vision algorithms, the team explains. The experts add that there are numerous ways in which the new technique could be used.

For starters, several industries could use it to analyze tiny components that are too large to fit under a conventional microscope. Further uses were identified in medicine (especially in-depth imaging), biometrics and forensics science.

According to the MIT experts, the key to the system is a slab of transparent, synthetic rubber called GelSight. One side of this slab is coated in a type of special paint that contains very small metal particles. The other side is transparent.

GelSight is applied to the object being studied, and then pressed on its surface. This makes the metal paint-coated side of the slab deform, and a camera is on stand-by to photograph the result. The algorithms then come into play, analyzing how the paint deformed.

The new system is an advanced version of a similar device that was tested back in 2009. The instrument was developed by the MIT John and Dorothy Wilson professor of vision science Edward Adelson, who is also a member of the Computer Science and Artificial Intelligence Laboratory.

He worked closely together with postdoctoral researcher Micah Kimo Johnson on developing the first version of GelSight. At the time, the system was able to detect the raised ink patterns on dollar bills.

The two experts then joined forces with MIT graduate student Alvin Raj and postdoctoral researcher Forrester Cole to improve at the technology, which was presented this year at the world's premier conference on computer graphics, Siggraph.

The improved GelSight system is able to register physical features less than a micrometer in depth and about two micrometers across, the group reported. Additionally, 3D representations of the studied objects can be reproduced, by shining light on the system from different angles.

Adelson says that the research team has already been contacted by forensics experts, who are very interested in the new technology for analyzing fingerprints.