Imaging Models
Imaging foundation models are trained on large-scale microscopy datasets — spanning fluorescence, cryo-electron microscopy, cell painting, and brightfield modalities — to learn visual representations that generalize across instruments, protocols, and specimen types. These models enable cell segmentation, image restoration, phenotypic profiling, and 3D reconstruction without task-specific retraining, accelerating workflows from high-content drug screening to structural biology. By capturing the shared visual grammar of biological images at scale, they reduce the annotation burden that has traditionally bottlenecked quantitative microscopy.
33 models in this category
Notable Models
Top-rated imaging models from our evaluations
BPD
Chan Zuckerberg Initiative
Fifth-place solution from the CZII CryoET Kaggle competition; an ensemble of four lightweight 3D U-Nets for protein particle localization in cryo-ET tomograms.
Cellpose 3
HHMI Janelia Research Campus
Generalist cell segmentation framework with a super-generalist cyto3 model and one-click image restoration networks optimized for downstream segmentation quality.
Cellpose-SAM
HHMI Janelia Research Campus
Generalist cell segmentation model combining SAM's ViT-L backbone with Cellpose flow fields. First model to surpass average human annotators on the Cellpose benchmark.
CryoLens
Chan Zuckerberg Initiative
A variational autoencoder for interpretable 3D reconstruction and representation learning of protein subtomograms from cryo-ET data, trained on 5.8 million synthetic particles.
Ensemble 3D UNet Soup
Chan Zuckerberg Initiative
Eighth-place CZII CryoET Kaggle solution; a weighted model soup of tiny, medium, and large 3D U-Nets pretrained on simulated data and fine-tuned on experimental cryo-ET tomograms.
MonjuDetectHM
Chan Zuckerberg Initiative
Seventh-place CZII CryoET Kaggle solution; an ensemble of three heatmap-predicting 3D segmentation models using ResNet50d and EfficientNetV2-M backbones for particle picking.