Stanford Scientists Develop Super Thin, High-Res Endoscope

Brian Buntz

March 13, 2013

2 Min Read
MDDI logo in a gray background | MDDI

Researchers at Stanford have created an endoscope with a diameter roughly as thin as a human hair. Capable of imaging objects 2.5 microns in size, the current prototype boasts a resolution four times greater than previous endoscope designs. In fact, the endoscope can produce four times as many image features than they expected.

The next-generation prototype may have a resolution as small as 0.3 microns, generating some 80,000 pixels. By contrast, conventional high-resolution endoscopes can resolve objects measuring approximately 10 microns.

The miniature endoscope illuminates objects using random intensity patterns (shown below) instead of a scanning spot (shown above). Image: Joseph Kahn, Stanford School of Engineering.

The endoscope, developed by Joseph Kahn, a Stanford professor of electrical engineering, could be potentially used for an array of medical applications, ranging from neurology to oncology.

The endoscope makes use of fibers in which light is split into numerous channels known as modes. While multimode fibers can convey a substantial amount of information, scrambling of the data travelling within them can be a problem. Kahn and his colleague Olav Solgaard,a also a professor of electrical engineering, were able to prevent the data from being scrambled by using a liquid crystal display known as a spatial light modulator.

In the past, multimode fibers generated images by forming a spot of light, which is scanned to a sequence of locations to sample an object (see the top part of the image on the right). The Stanford researchers used random speckle patterns for imaging, and used  linear optimization to reconstruct the images.

The main limitation of the endoscope is that it must remain rigid, as flexing it scrambles the images it generates.

The research was published in Optics Express.

Related Content

Sign up for the QMED & MD+DI Daily newsletter.

You May Also Like