Cellular segmentation is an important tool in quantitative cell biology that is often required when measuring properties such as cell shape, position, or RNA/protein expression. This task is time consuming when done manually and can be quite difficult to automate as the variety of microscopy modalities, markers, and cell types are hard generalize. To improve this generalizability, Carson Stringer and colleagues have developed Cellpose, a generalist algorithm for cell and nucleus segmentation.
Cellpose is a deep learning-based segmentation method and GUI that was trained on a new, highly diverse image set. The image set includes the novel addition of non-cellular images containing repeated objects such as fruit and jellyfish that are expected to improve generalizability. The tool significantly outperforms other segmentation methods on several benchmarks and due to community involvement may improve further with the addition of user contributed training data. Cellpose can also be used for both 2D and 3D cell segmentation. Cellpose is a powerful cellular segmentation algorithm that implements novel techniques for improved precision and generalizability.
This research tool was created by your colleagues. Please acknowledge the Principal Investigator, cite the article in which the tool was described, and include an RRID in the Materials and Methods of your future publications. RRID:SCR_022332