Cellpose-SAM: superhuman generalization for cellular segmentation
Pachitariu, M.; Rariden, M.; Stringer, C.
10.1101/2025.04.28.651001 bioRxivShow abstract
Modern algorithms for biological segmentation can match inter-human agreement in annotation quality. This however is not a performance bound: a hypothetical human-consensus segmentation could reduce error rates in half. To obtain a model that generalizes better we adapted the pretrained transformer backbone of a foundation model (SAM) to the Cellpose framework. The resulting Cellpose-SAM model substantially outperforms inter-human agreement and approaches the human-consensus bound. We increase generalization performance further by making the model robust to channel shuffling, cell size, shot noise, downsampling, isotropic and anisotropic blur. The new model can be readily adopted into the Cellpose ecosystem which includes finetuning, human-in-the-loop training, image restoration and 3D segmentation approaches. These properties establish Cellpose-SAM as a foundation model for biological segmentation.
Matching journals
The top 3 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Generative AI Enables Medical Image Segmentation in Ultra Low-Data Regimes 96%
- Large-scale capture of hidden fluorescent labels for training generalizable markerless motion capture models 96%
- Segmenting functional tissue units across human organs using community-driven development of generalizable machine learning algorithms 94%
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- Whole-cell segmentation of tissue images with human-level performance using large-scale data annotation and deep learning 95%
- Automated Reconstruction of Whole-Embryo Cell Lineages by Learning from Sparse Annotations 95%
- scvi-tools: a library for deep probabilistic analysis of single-cell omics data 94%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.