AI has improved advanced microscopy

Scientists have combined advanced microscopy techniques and AI-based algorithms to observe fast biological processes in 3D. This data is processed within a few seconds.

To observe the rapid signals from neurons in the fish brain, scientists have begun using a technique called brightfield. It provides 3D images of fast biological processes. However, image quality is often poor, and it takes several days on average to convert huge amounts of data into 3D volumes.

Now scientists from the European Molecular Biology Laboratory (EMBL) have combined artificial intelligence (AI) algorithms with two advanced microscopy techniques to reduce image processing time from days to seconds while maintaining image clarity and accuracy. The results of the study appeared in the journal Nature Methods.

“In this development, we were able to combine the best of both worlds,” says Niels Wagner, one of the two lead authors of the article and a graduate student at the Technical University of Munich. “AI has allowed us to combine different microscopy techniques so that we can acquire images very quickly without compromising on quality.”

To take advantage of each approach, EMBL researchers have developed an approach that uses brightfield microscopy to image large 3D specimens and plane illumination microscopy to train AI algorithms, which then create an accurate 3D image of the specimen.

“If you imagine algorithms that create an image, you need to verify that those algorithms are producing the correct image,” explains Anna Kreshuk, EMBL team leader, whose team brought machine learning expertise to the project. “In the new study, scientists used light-sheet microscopy to make sure the AI ​​algorithms were working. This distinguishes our research from what has been done in the past. “

Author: John Kessler
Graduated From the Massachusetts Institute of Technology. Previously, worked in various little-known media. Currently is an expert, editor and developer of Free News.
Function: Director
127 number 0.857912 time