A new AI tool developed by the University of Gothenburg offers new possibilities for analyzing images taken with microscopes. The tool has been recognized internationally and is now used in industry and research.
This study focuses on our daily use of deep learning, AI, and machine education. When a song pops up on Spotify similar to one we’ve listened to before, or when the mobile camera automatically detects the best settings for a photo and corrects color.
Deep learning has made a significant impact in many sectors and industries. “We have developed a tool that allows deep learning to be used with an emphasis on images taken using microscopes,” Benjamin Midtvedt (a doctoral student in Physics and the study’s principal author) says.
Deep learning is a mathematical model that solves impossible problems using traditional algorithmic methods. Deep learning has been proven very effective in microscopy.
Midtvedt’s research team has developed a neural network learning tool that can extract the exact information a researcher needs from an image. This is done by looking through large numbers of images (known as training data). This tool makes it easier to produce training data than manually. It can generate thousands of pictures in one hour instead of hundreds a month.
This allows you to quickly extract more detail from microscope images without performing a complex analysis using traditional methods. The results can also be customized to retrieve specific information for specific purposes.
The tool can determine the size and characteristics of microscopic particles and count and classify cells. Researchers have demonstrated that this tool can be used in industries that want to purify their emissions. They can check in real-time whether any unwanted particles have been removed.
Researchers hope that the tool will be able to track infections within cells and map cell defense mechanisms in the future. This would lead to enormous opportunities for new treatments and medicines. Researchers can conduct more analyses, make discoveries and implement new ideas, regardless of the microscopical challenges.