How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!
"Making direct measurements of the hippocampus and entorhinal cortex can be very difficult," Bartha explains. "It meant manually tracing many slices in the brain. It was tedious and prone to error because a lot of judgment was involved.
More recently, Nick Fox actually developed software that registers brain scans acquired at different times and based on how much warping is required can measure rates of atrophy. He showed that the rate of whole brain atrophy was 3 to 5 percent over a year in patients with Alzheimer's disease."
The Robarts researchers use software from Cedara Software, Mississauga, Ontario. They work with a team headed by Vittorio Accomazzi, Lead Architect Image Processing at Cedara. "We joined forces two or three years ago," Bartha says. "We met for a different reason, but it gave me a chance to see some of Vittorio's software and how it could be adapted to brain ventricles. We essentially provide ideas to develop it and Vittorio's group produced software for us. We give them feedback and they very quickly provide new prototypes."
The technology revolves around segmentation, or portioning the image into multiple regions, making it easier to analyze. Cedara actually has been using the technology for other programs, including segmenting of aortas or kidneys.
"We saw an opportunity to deploy the technology we've been using," Accomazzi explains. "It's called Fuzzy Region Growing and it's the technology used to identify the ventricles on the MR images. Basically, it's a method to deal with uncertainty. They click on the point or ventricle and the system identifies the entire ventricle. The application Robarts is using is leveraging this technology and is optimized to minimize user intervention, to make the measurements easier, faster and operator independent. It takes the human element out of the measurement."