Measure Bioluminescent Signals 10 Times More Precisely with AI

KAIST Publishes 'Nature Methods' Paper on the 19th

KAIST announced on the 20th that a research team led by Professor Youngkyu Yoon from the Department of Electrical Engineering has developed an artificial intelligence (AI) image analysis technology that enables measurement of biological fluorescence signals with more than 10 times higher precision compared to existing technologies.

Conceptual diagram of support technology. Image source: Provided by KAIST

Recently, with advancements in genetic engineering technology, techniques have been developed and utilized that use fluorescence microscopy to continuously capture and measure signals within living biological tissues by converting them into fluorescence signals. However, because the fluorescence signals emitted from biological tissues are weak, when measuring rapidly changing signals such as electrical signals of nerve cells, the signal-to-noise ratio becomes very low, making precise measurement difficult.

The research team developed a technology that, without any separate training data, autonomously learns the statistical distribution of data from fluorescence microscopy images with low signal-to-noise ratios, enhancing the signal-to-noise ratio of the images by more than 10 times to enable precise measurement of biological signals. By applying this technology, the measurement precision of various biological signals can be significantly improved, which is expected to be widely utilized in life science research overall and in the development of treatments for brain diseases.

Professor Yoon stated, “This technology can be easily used by researchers employing various fluorescence imaging devices without the need for separate training data, and it is expected to be broadly applied to uncover new biological phenomena.”

The research results were published online on the 19th in the international academic journal Nature Methods. It was also selected as the cover paper for the October issue.

This content was produced with the assistance of AI translation services.