How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!
JEOL USA announced that researchers at Harvard University's Department of Molecular and Cellular Biology have selected JEOL as a partner in a collaborative effort to map the brain using high-resolution SEM images. Harvard biologist Jeff Lichtman, post-doctorate student Narayanan (Bobby) Kasthuri, and University of Southern California research assistant Kenneth Hayworth plan to use a JEOL scanning electron microscope (SEM) to ultimately produce a 3D image of the entire mouse brain. Harvard took delivery of a JEOL high-resolution field emission SEM in June 2007.
Working together, the JEOL and Harvard teams will customize the JSM-7001FLV, one of JEOL's most popular SEMs used for advanced imaging and research, to sequentially allow fast acquisition of hundreds of thousands of high pixel density images at the nanometer scale. "We'll ultimately be using image recognition software for montaging at very high pixel resolution—approximately five nanometers—to acquire up to 20 megabyte images in about three seconds," said Charles Nielsen, JEOL USA SEM product manager and vice president.
"We've learned a lot about SEM and robotics," said Kasthuri, who said the Harvard lab team chose to partner with JEOL both for the SEM's capabilities and the help of the JEOL team.
To produce the sample images, the Lichtman lab at Harvard is developing a new device that will automatically slice the sample and embed the organelles into a resin for sequential imaging. Hayworth, Kasthuri and Lichtman will head development and utilization of the Automatic Tape-Collecting Lathe-Ultramicrotome designed to produce ultra thin (50nm and less) slices of the brain, and retain them in sequential order for reconstruction mapping. The entire sample, typically measuring 1mm x 1mm, will be embedded into a half-mile long substrate.
"We hope to eventually do cutting and data collection without human intervention," said Kasthuri. "Not that it couldn't have been done 30 years ago, but we couldn't accumulate such large amounts of data before there was digital imaging."