AI taught to explore living cells

Researchers at the University of Illinois (USA) have developed a new technique that combines visualization with

artificial intelligence (AI) to study unlabeled living cells over a long period of time. This technique can be used to investigate cell viability and pathology.

“Our laboratory specializes inimages without additional labels, which allows us to visualize cells without the use of toxic chemicals, - said Gabriel Popescu, professor of electrical and computer engineering, director of the Laboratory for Quantitative Imaging of Light at the Beckmann Institute of Advanced Science and Technology. “However, we cannot measure specific properties of a cell without the use of toxic fluorescent dyes. We solved this problem in our research. "

Therefore, scientists have used computational methods that can follow the shape and condition of the sample without killing the cells.

Sberbank has developed a platform for working with AI

At first, the researchers followed the cells infor several days using its non-destructive, label-free technique. At the end of the experiment, they colored the samples and used deep learning to find out where the fluorescent dyes would be.

Despite the fact that previously artificial intelligencewas used to create one type of image, researchers were able to program it to analyze photographs in real time. They were able to study the cells to pinpoint exactly where the parts are.

Another advantage of this technique isthat scientists can conduct experiments for many days. The cells remain alive even after a week. This cannot be done with fluorescent dyes, as chemical toxicity can kill cells. Researchers are now trying to adapt algorithms to study various properties of cells and biological samples.

Read also:

The Doomsday glacier turned out to be more dangerous than scientists thought. We tell the main thing

See the largest iceberg ever taken from the air

China first tested "artificial sun"