U-M researcher shoots to speed image and video processing 1,000x
ANN ARBOR, Mich. — A University of Michigan researcher is leading a project to build computer hardware that would process images and video 1,000 times faster than current systems.
The Ann Arbor school announced this week that Wei Lu, an associate professor of electrical engineering and computer science, has been awarded a contract of up to $5.7 million from the Defense Advanced Research Projects Agency. So far, Lu has received $1.3 million.
The project is titled "Sparse Adaptive Local Learning for Sensing and Analytics." It includes designing and fabricating a computer chip to do the work.
Rather than rendering pixel by pixel, as today's computers do, new computer systems will look at a whole image at once and identify structures in the image.