Cellpose: a generalist algorithm for cellular segmentation
By C. Stringer, Tim Wang, Michalis Michaelos, Marius Pachitariu. Nature Methods, 2020
3,569 citations recorded by the source corpus.
Many biological applications require the segmentation of cell bodies, membranes and nuclei from microscopy images. Deep learning has enabled great progress on this problem, but current methods are specialized for images that have large training datasets. Here we introduce a generalist, deep learning-based segmentation method called Cellpose, which can precisely segment cells from a wide range of image types and does not require model retraining or parameter adjustments. Cellpose was trained on a new dataset of highly varied images of cells, containing over 70,000 segmented objects. We also demonstrate a three-dimensional (3D) extension of Cellpose that reuses the two-dimensional (2D) model and does not require 3D-labeled data. To support community contributions to the training data, we developed software for manual labeling and for curation of the automated results. Periodically retraining the model on the community-contributed data will ensure that Cellpose improves constantly. Cellpose is a generalist, deep learning-based approach for segmenting structures in a wide range of image types. Cellpose does not require parameter adjustment or model retraining and outperforms established methods on 2D and 3D datasets.
Subfields: Biophysics, Media Technology, Molecular Biology
View Paper: https://www.semanticscholar.org/paper/8f1a8b82c7be223f195b4f03ffa1943391fd428b
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Biophysics
- C. Stringer
- Marius Pachitariu
- Malte D. Luecken
- Loic A. Royer
- Florian Jug
- Alexander Krull
- Tim-Oliver Buchholz
- Luke Zappia
In Media Technology
- Marius Pachitariu
- C. Stringer
- Loic A. Royer
- Florian Jug
- Alexander Krull
- Tim-Oliver Buchholz
- Anne E Carpenter
- Fabian J Theis
In Molecular Biology
- Malte D. Luecken
- Fabian J Theis
- C. Stringer
- Marius Pachitariu
- Luke Zappia
- H. Schiller
- Anna C. Schaar
- F. Drost
Citations in the corpus
- CellProfiler 4: improvements in speed, utility and usability (2021, methodology): “For CellProfiler 3 we previously released a plugin for NucleAIzer [15], and in the future we hope to investigate integrations with other popular models such as Cellpose [16] and Stardist [17].”
- Best practices for single-cell analysis across modalities (2023, methodology): “…, for example, leverage spatial correlation testing, BayesSpace 227 uses Markov random fields, spaGCN 229 uses graph neural networks to integrate gene expression data, spatial information and histology images, and sepal 235 utilizes diffusion-based modelling to identify genes with spatial patterns.”
- Metrics Reloaded: Recommendations for image analysis validation (2022, neutral): “Despite these problems, the terminology has been adapted by subsequent influential work [49, 76, 82] indicating a wide-spread usage within the community.”
- Cellpose 2.0: how to train your own model (2022, methodology): “The original Cellpose is a generalist model that can segment a wide variety of cellular images [5].”
- TrackMate 7: integrating state-of-the-art segmentation algorithms into tracking pipelines (2022, methodology): “Mouse hematopoietic stem cells migrating in a hydrogel microwell 17 were automatically segmented using cellpose (Cyto model) implemented in the ZeroCostDL4Mic platform 12,15 .”
- Democratising deep learning for microscopy with ZeroCostDL4Mic (2021, methodology): “Using pretrained models alleviates the computational requirement of training and has therefore become popular for tasks such as cellular and nuclear segmentation 6,17,22 .”
- Cellpose3: one-click image restoration for improved cellular segmentation (2024, positive): “In addition, we employ perceptual loss functions to make the reconstruction perceptually similar to the clean images [18, 19], and we use a balanced combination of large-scale datasets for training [20–26].”
- Labkit: Labeling and Segmentation Toolkit for Big Image Data (2021, background): “This method was designed as part of a cell segmentation and tracking pipeline on various images, and it is likely that recent and specialized deep learning segmentation algorithms, such as StarDist (Schmidt et al., 2018) or CellPose (Stringer et al., 2021), would perform overall better (Baltissen et”
- High resolution mapping of the tumor microenvironment using integrated single-cell, spatial and in situ analysis (2023, neutral): “This approach is conceptually similar to the popular CellPose algorithm 36 in that an encoder-decoder neural network does not directly solve for the segmentation mask, but instead solves a related problem that is easier to learn and allows for the imposition of geometric constraints.”