The team at IFOM conducts numerous experiments on yeast cells, where it is important to accurately track cell movement, as this reveals a lot about the cells themselves. Although there are already several solutions available that automate this task, none of the existing methods are completely accurate. Our researchers — Gergely Szabó, Andrea Ciliberto, and András Horváth — aimed to improve this by increasing the accuracy of cell movement tracking.

They introduced two key methodological innovations.

First, unlike previous approaches, they utilized both future and past information from the microscopic samples to more precisely identify the cells’ paths.

As a second innovation, they made use of subtle features in the imaging data that help infer the likely direction of a cell’s movement. They trained the artificial intelligence model to predict the likely location of a given cell in the next frame based on its prior movement data.

You can find more detailed information about the article at this link.