Could artificial intelligence help scientists fight toxic blooms in the Great Lakes?
We’re coming up on the time of year when people will be testing lakes for toxic blooms of cyanobacteria.
Jason Deglint wants to speed up that testing process. Right now, he says it can take at least a few days.
“The normal procedure is they take a water sample and they have to ship it to a certified lab," Deglint said. "Then a highly trained professional will sit down at a microscope, and they’ll manually identify and count these different organisms."
Deglint is getting his PhD from the University of Waterloo. He and his colleagues have built a prototype of a custom microscope system. First, he says, they take a digital image of a water sample, and then teach a computer to quickly identify algae or cyanobacteria in that sample.
"Just like when you teach a child to recognize things, first you teach it different types of cars, then it starts to know the difference between a car and a truck; we're showing examples of different types of organisms and species to our computer," Deglint explained. "We’re teaching it what these look like so when it sees one that it’s never seen before, it can now identify it."
Deglint says the goal is to be able to read a water sample within minutes.
He says the prototype will take another year to finish. The next step is to test the computer against human experts in the lab.