Â鶹ÒùÔº

Picking out individual cells in distorted microscopy images is now as easy as clicking a button.

A new version of —the popular tool that maps the boundaries of diverse cells in microscopy images—now works on less-than-perfect pictures that are noisy, blurry, or undersampled.

Many general methods used for segmenting individual cells in microscopy images, including previous versions of Cellpose, have a hard time recognizing cellular boundaries that have been distorted by noise, blurring, or undersampling.

Janelia Group Leaders Carsen Stringer and Marius Pachitariu set out to address this issue with the development of Cellpose3. Unlike previous approaches, which train models to improve the quality of distorted images, Cellpose3 was instead trained to improve the segmentation of distorted images.

The study is in the journal Nature Methods.

The Cellpose3 restoration algorithm, when applied to distorted images, produces crisp restored images which can then be easily segmented by the original Cellpose segmentation algorithm.

Cellpose3 is also trained on a large, varied collection of images, enabling users to easily use the new method, which is available as a "one-click" button in the Cellpose application, on their own data.