cellpose

cellpose is an anatomical segmentation algorithm written in Python 3 by Carsen Stringer and Marius Pachitariu. For support, please open an issue.

We make pip installable releases of cellpose, here is the pypi. You can install it as pip install cellpose[gui].

You can try it out without installing at cellpose.org. Also check out these resources:

Cellpose3: one-click image restoration for improved cellular segmentation

• paper on biorxiv

• thread

Cellpose 2.0: how to train your own model

• paper on biorxiv

• talk

• twitter thread

• human-in-the-loop training protocol video

Cellpose: a generalist algorithm for cellular segmentation

• paper on biorxiv (see figure 1 below) and in nature methods

• twitter thread

• Marius’s talk

Basics:

• Installation
  • Built-in model directory
  • M1-M3 Mac installation
  • AMD GPU ROCm installation
  • Common issues
  • Dependencies
• GUI
  • Starting the GUI
  • Using the GUI
  • Drawing masks
  • Bulk Mask Deletion
  • Segmentation options
  • Training your own cellpose model
  • Contributing training data
  • Keyboard shortcuts
• Inputs
• Settings
  • Channels
  • Diameter
  • Resample
  • Flow threshold
  • Cellprob threshold
  • Number of iterations niter
• Outputs
  • in a notebook
  • _seg.npy output
  • PNG output
  • Native ImageJ ROI archive output
  • (Legacy ImageJ Interface) ROI manager compatible output for ImageJ
  • Plotting functions
• 3D segmentation
  • Input format
  • Segmentation settings
  • Training for 3D segmentation
• Models
  • Full built-in models
    • Cytoplasm model ('cyto3', 'cyto2', 'cyto')
    • Nucleus model (‘nuclei’)
  • Other built-in models
  • User-trained models
    • Finding models on BioImage.IO
    • Sharing models on BioImage.IO
• Image Restoration
  • DenoiseModel
  • CellposeDenoiseModel
  • Command line usage
  • Training your own models
• Training
  • CLI training options
  • Re-training a model
• Timing + memory usage
• Big Data
  • Distributed Cellpose for larger-than-memory data
  • Examples
• OpenVINO
• FAQ

Examples:

• In a notebook
• Command line
  • Command Line Usage

Reference:

• Cellpose API Guide
  • Cellpose class
    • Cellpose
      • Cellpose.device
      • Cellpose.gpu
      • Cellpose.diam_mean
      • Cellpose.cp
      • Cellpose.pretrained_size
      • Cellpose.sz
      • Cellpose.eval()
  • CellposeModel
    • CellposeModel
      • CellposeModel.diam_mean
      • CellposeModel.builtin
      • CellposeModel.device
      • CellposeModel.mkldnn
      • CellposeModel.nchan
      • CellposeModel.nclasses
      • CellposeModel.nbase
      • CellposeModel.net
      • CellposeModel.pretrained_model
      • CellposeModel.pretrained_model_ortho
      • CellposeModel.backbone
      • CellposeModel.__init__()
      • CellposeModel.eval()
      • CellposeModel.eval()
  • CellposeDenoiseModel
    • CellposeDenoiseModel
      • CellposeDenoiseModel.eval()
  • DenoiseModel
    • DenoiseModel
      • DenoiseModel.nchan
      • DenoiseModel.diam_mean
      • DenoiseModel.net
      • DenoiseModel.pretrained_model
      • DenoiseModel.net_chan2
      • DenoiseModel.net_type
      • DenoiseModel.eval()
  • SizeModel
    • SizeModel
      • SizeModel.pretrained_size
      • SizeModel.cp
      • SizeModel.device
      • SizeModel.diam_mean
      • SizeModel.eval()
  • Training
    • train_seg()
    • train_size()
  • Metrics
    • aggregated_jaccard_index()
    • average_precision()
    • boundary_scores()
    • flow_error()
    • mask_ious()
  • Flows to masks
    • compute_masks()
    • follow_flows()
    • get_centers()
    • get_masks_orig()
    • get_masks_torch()
    • labels_to_flows()
    • map_coordinates()
    • masks_to_flows()
    • masks_to_flows_cpu()
    • masks_to_flows_gpu()
    • masks_to_flows_gpu_3d()
    • max_pool1d()
    • max_pool2d()
    • max_pool3d()
    • max_pool_nd()
    • remove_bad_flow_masks()
    • resize_and_compute_masks()
    • steps2D()
    • steps3D()
    • steps_interp()
  • Image transforms
    • average_tiles()
    • convert_image()
    • gaussian_kernel()
    • make_tiles()
    • move_axis()
    • move_min_dim()
    • normalize99()
    • normalize99_tile()
    • normalize_img()
    • pad_image_ND()
    • random_rotate_and_resize()
    • reshape()
    • resize_image()
    • resize_safe()
    • smooth_sharpen_img()
    • unaugment_tiles()
    • update_axis()
  • Plot functions
    • disk()
    • dx_to_circ()
    • image_to_rgb()
    • interesting_patch()
    • mask_overlay()
    • mask_rgb()
    • outline_view()
    • show_segmentation()
  • I/O functions
    • add_model()
    • get_image_files()
    • get_label_files()
    • imread()
    • imsave()
    • load_images_labels()
    • load_train_test_data()
    • masks_flows_to_seg()
    • remove_model()
    • save_masks()
    • save_rois()
    • save_to_png()
  • Utils functions
    • TqdmToLogger
      • TqdmToLogger.flush()
      • TqdmToLogger.write()
    • circleMask()
    • diameters()
    • dilate_masks()
    • distance_to_boundary()
    • download_url_to_file()
    • fill_holes_and_remove_small_masks()
    • get_mask_compactness()
    • get_mask_perimeters()
    • get_mask_stats()
    • get_masks_unet()
    • get_outline_multi()
    • get_perimeter()
    • masks_to_edges()
    • masks_to_outlines()
    • outlines_list()
    • outlines_list_multi()
    • outlines_list_single()
    • radius_distribution()
    • remove_edge_masks()
    • size_distribution()
    • stitch3D()
  • Network classes
    • CPnet
      • CPnet.nbase
      • CPnet.nout
      • CPnet.sz
      • CPnet.residual_on
      • CPnet.style_on
      • CPnet.concatenation
      • CPnet.conv_3D
      • CPnet.mkldnn
      • CPnet.downsample
      • CPnet.upsample
      • CPnet.make_style
      • CPnet.output
      • CPnet.diam_mean
      • CPnet.diam_labels
      • CPnet.device
      • CPnet.forward()
      • CPnet.load_model()
      • CPnet.save_model()
    • CPnetBioImageIO
      • CPnetBioImageIO.forward()
      • CPnetBioImageIO.load_model()
      • CPnetBioImageIO.load_state_dict()
    • batchconvstyle
    • downsample
    • make_style
    • resdown
    • resup
    • upsample
  • Core functions
    • assign_device()
    • check_mkl()
    • run_3D()
    • run_net()
    • use_gpu()
  • All models functions
    • Cellpose
      • Cellpose.device
      • Cellpose.gpu
      • Cellpose.diam_mean
      • Cellpose.cp
      • Cellpose.pretrained_size
      • Cellpose.sz
      • Cellpose.eval()
    • CellposeModel
      • CellposeModel.diam_mean
      • CellposeModel.builtin
      • CellposeModel.device
      • CellposeModel.mkldnn
      • CellposeModel.nchan
      • CellposeModel.nclasses
      • CellposeModel.nbase
      • CellposeModel.net
      • CellposeModel.pretrained_model
      • CellposeModel.pretrained_model_ortho
      • CellposeModel.backbone
      • CellposeModel.__init__()
      • CellposeModel.eval()
      • CellposeModel.eval()
    • SizeModel
      • SizeModel.pretrained_size
      • SizeModel.cp
      • SizeModel.device
      • SizeModel.diam_mean
      • SizeModel.eval()
    • get_model_params()
• Cellpose CLI
  • Command Line Usage
    • Named Arguments
    • Hardware Arguments
    • Input Image Arguments
    • Model Arguments
    • Algorithm Arguments
    • Output Arguments
    • Training Arguments