cellpose - v4.0.5

Updates to Cellpose-SAM

• resample added back to CLI for diameter changes. Note that the diameter parameter is only used for image rescaling and is not always necessary to provide in CPSAM compared to CP3.
• default model weights are changed to torch.bfloat16 instead of previous 32-bit float data type. This decreases the model size by ~50% and speeds up runtime by ~40%, therefore helping with OOM issues. Changing back to the 32-bit model can be done during CellposeModel instantiation by setting use_bfloat16=False.
• Improved 3d image dimension guessing
• Fixed a bug that ran the network 4x more often than necessary when image size was <256 length or width.
• Added some documentation
• Updated notebooks
• Some more comprehensive testing

Changelog: https://github.com/MouseLand/cellpose/compare/v4.0.4...v4.0.5

- Python
Published by mrariden 8 months ago

cellpose - v4.0.4

Introducing Cellpose-SAM!

New in cellpose v4

• Cellpose-SAM aka CP4 is fully integrated in to the CLI/API/GUI and notebooks
• cellpose.models.Cellpose class was removed, users should now only use cellpose.models.CellposeModel
• Size estimation via cellpose.models.SizeModel class was removed. Users don't have to use the diameter setting since CP4 is much more robust to object size.
• Denoising components via cellpose.denoise is currently not available.
• Added more comprehensive testing
• Better utilization of MPS for Mac users when doing 3D segmentation
• Added --save_each flag for training from the CLI
• More details are available in the paper, talk, and in the docs.

New Contributors

• @KevinCortacero made their first contribution in https://github.com/MouseLand/cellpose/pull/1080
• @unikill066 made their first contribution in https://github.com/MouseLand/cellpose/pull/1159

Changelog for early cellpose v4 changes: https://github.com/MouseLand/cellpose/compare/v4.0.1...v4.0.4

- Python
Published by mrariden 9 months ago

cellpose - v3.1.1.2

This is the last release of Cellpose3 before the Cellpose-SAM release.

What's Changed

• fix tiling bug when augment=True
• Documented using masks and preprocessing_steps in distributed rst by @GFleishman in https://github.com/MouseLand/cellpose/pull/1108
• Compress TIFFs on save by @qin-yu in https://github.com/MouseLand/cellpose/pull/1131
• Report errors when saving masks to *_seg.npy by @deroulers in https://github.com/MouseLand/cellpose/pull/1129
• Correct GUI initialization order of load_3D for modular usage. by @biswajitSM in https://github.com/MouseLand/cellpose/pull/1119
• Optimize fillholesandremovesmall_masks by @Tomvl117 in https://github.com/MouseLand/cellpose/pull/1116
• pyarrow missing by @GFleishman in https://github.com/MouseLand/cellpose/pull/1109

New Contributors

• @biswajitSM made their first contribution in https://github.com/MouseLand/cellpose/pull/1119
• @Tomvl117 made their first contribution in https://github.com/MouseLand/cellpose/pull/1116

Full Changelog: https://github.com/MouseLand/cellpose/compare/v3.1.1.1...v3.1.1.2

- Python
Published by carsen-stringer 10 months ago

cellpose - v3.1.1

New distributed cellpose module enables running of cellpose on really big data, on workstations and clusters, thanks to the heroic work of @GFleishman ! Check out the documentation here.

We made bug fixes for Mac from v3.1 and other small bug fixes, thanks to all the contributors.

What's Changed

• fix bug with channels=None in CellposeDenoiseModel (#1098)
• add z_axis in CellposeDenoiseModel (#1099)
• to avoid MPS errors, use CPU for Mac in get_masks_torch and use float32 in _extend_centers_gpu (#1034, #1063, #1097)
• fix now masks found issue (#1038)
• add weights_only flag (#1042)
• bug with invert=True (#1056)
• fix issue with version splitting (#1074)
• fix bug with anisotropy in make_train (#1076)
• fixing issue when channels and z are equal in stack (#1086)
• Larger than memory images: performant and scalable distributed implementation for workstations and clusters by @GFleishman in https://github.com/MouseLand/cellpose/pull/1062
• Bugfix: Removing unused 'ARBITRARY_SIZE' by @Pablo1990 in https://github.com/MouseLand/cellpose/pull/1082
• train.py: remove superfluous percent sign in f-string by @deroulers in https://github.com/MouseLand/cellpose/pull/1077
• Add norm percentiles as CLI argument by @TheBestDestroyer in https://github.com/MouseLand/cellpose/pull/1058
• [Doc] Describe how to manage multithreading by @royjacobson in https://github.com/MouseLand/cellpose/pull/1045
• Fix saving of flow fields and cell probability in GUI by @qin-yu in https://github.com/MouseLand/cellpose/pull/1103
• File selection cancel logic added by @knowblesse in https://github.com/MouseLand/cellpose/pull/1040
• feat: custom output name by @antoinemeyer5 in https://github.com/MouseLand/cellpose/pull/1104

New Contributors

• @GFleishman made their first contribution in https://github.com/MouseLand/cellpose/pull/1062
• @Pablo1990 made their first contribution in https://github.com/MouseLand/cellpose/pull/1082
• @deroulers made their first contribution in https://github.com/MouseLand/cellpose/pull/1077
• @TheBestDestroyer made their first contribution in https://github.com/MouseLand/cellpose/pull/1058
• @royjacobson made their first contribution in https://github.com/MouseLand/cellpose/pull/1045
• @knowblesse made their first contribution in https://github.com/MouseLand/cellpose/pull/1040
• @antoinemeyer5 made their first contribution in https://github.com/MouseLand/cellpose/pull/1104

Full Changelog: https://github.com/MouseLand/cellpose/compare/v3.1.0...v3.1.1

- Python
Published by carsen-stringer about 1 year ago

cellpose - v3.1.0

Check out the additional options for performing 3D segmenation, and more guidance here. Also, the latest mask creation function get_masks_torch reduces CPU RAM usage compared to the previous version, and is available to run on the GPU for faster mask creation.

Updated info with CPU/GPU RAM usage and approx runtimes for single images, available here.

Also we should have full compatibility with MPS at least in 2D (thanks @OratHelm !) And thanks to @qin-yu you can share your cellpose models to the bioimage.io model zoo.

What's Changed

• MPS compatibility on Apple Silicon by @OratHelm in https://github.com/MouseLand/cellpose/pull/1003
• Rev3 by @carsen-stringer, adds additional denoising models and anisotropy models, in https://github.com/MouseLand/cellpose/pull/1008
• Add missing train/test files to processtrain_test() call by @kemaleren in https://github.com/MouseLand/cellpose/pull/995
• imagecodecs no longer compiles with python 3.8 by @ajinkya-kulkarni in https://github.com/MouseLand/cellpose/pull/979
• Fix link in doc by @maweigert in https://github.com/MouseLand/cellpose/pull/970
• allows --all channels to work in CLI by @lacan in https://github.com/MouseLand/cellpose/pull/949
• Export CPnet to BioImage.IO Model Zoo by @qin-yu in https://github.com/MouseLand/cellpose/pull/988
• Bugfix: Resizing masks containing over 65,535 cells not supported by OpenCV by @Tobiaspk in https://github.com/MouseLand/cellpose/pull/938
• Add .flex to the allowed image file extensions by @mmasinas in https://github.com/MouseLand/cellpose/pull/881
• added optional params cppath and logfilename to the setup_logger fu… by @MaxJoas in https://github.com/MouseLand/cellpose/pull/862
• Skip empty outlines and display a message in the console. by @francodgstn in https://github.com/MouseLand/cellpose/pull/829
• Fix intensity normalizations by @qin-yu in https://github.com/MouseLand/cellpose/pull/981
• Add docs for BioImage.IO: finding & sharing models by @qin-yu in https://github.com/MouseLand/cellpose/pull/1011
• New3d by @carsen-stringer, adds new mask creation script on GPU, in https://github.com/MouseLand/cellpose/pull/1030
• GUI selection for SGD/RAdam Training by @ian-coccimiglio in https://github.com/MouseLand/cellpose/pull/1022

New Contributors

• @OratHelm made their first contribution in https://github.com/MouseLand/cellpose/pull/1003
• @kemaleren made their first contribution in https://github.com/MouseLand/cellpose/pull/995
• @maweigert made their first contribution in https://github.com/MouseLand/cellpose/pull/970
• @lacan made their first contribution in https://github.com/MouseLand/cellpose/pull/949
• @qin-yu made their first contribution in https://github.com/MouseLand/cellpose/pull/988
• @Tobiaspk made their first contribution in https://github.com/MouseLand/cellpose/pull/938
• @mmasinas made their first contribution in https://github.com/MouseLand/cellpose/pull/881
• @MaxJoas made their first contribution in https://github.com/MouseLand/cellpose/pull/862
• @ian-coccimiglio made their first contribution in https://github.com/MouseLand/cellpose/pull/1022

Full Changelog: https://github.com/MouseLand/cellpose/compare/v3.0.10...v3.1.0

- Python
Published by carsen-stringer over 1 year ago

cellpose - Cellpose v3.0.10

• fix bug in diameter changing when rerunning the model after training -- should make Cellpose 3 training parameters now equivalent to Cellpose 2
• fix bug with nimgperepoch being fixed to 8 when training in the GUI (#925)

- Python
Published by carsen-stringer over 1 year ago

cellpose - Cellpose v3.0.8

• Update install instructions to include venv
• Documentation fixes and additions
• Small bug fixes
• fix io.masks_flows_to_seg() iteration bug

- Python
Published by mrariden almost 2 years ago

cellpose - Cellpose v3.0.5

Cellpose3 release with CLI, API and GUI support! Check out docs for updates to code: https://cellpose.readthedocs.io/en/latest/restore.html

- Python
Published by carsen-stringer about 2 years ago

cellpose - Cellpose v3.0.1

first Cellpose 3 release!

- Python
Published by carsen-stringer about 2 years ago

cellpose - cellpose v2.3.2

Last release of Cellpose 2 before update to Cellpose 3.

• fix slow stitching problem by only checking ROIs for stitching in current plane (issue #845)
• fix reshaping issue in flows from @mrariden ('mrariden-dynamicsreshapefix' )
• add python 3.11 support from @harryzcy (#807)
• disable multiproc on windows from @mrariden
• fixes bug in 3D interpolation for multi-ROI masks
• disable gui autosave from @mrariden (#784)
• fix min size from @SimonBon (#781)
• fix imagejroiconverter from @francodgstn (#828)
• fix torch version command (#831)
• fix zaxis=0 bug from @nickeener (#688)
• add a pyqt version to env from @rhoadesScholar (#849)
• fix for multimodel adding to gui from @landoskape (#839)

- Python
Published by carsen-stringer about 2 years ago

cellpose - cellpose v2.2.2

Changes added to main branch

• Added functionality to allow saving ROIs in ImageJ-native format.
• Minor bug fix related to saving images with many objects thanks to @alix-simon
• Performance improvement using improved indexing thanks to @sophiamaedler
• Update Read the Docs to include a FAQ section, and misc RTD updates
• Add multithreading to 'outlines' exporting to speed up saving/execution time

- Python
Published by mrariden almost 3 years ago

cellpose - cellpose v2.2.1

• fixing deprecation of np.bool
• @mrariden improved docs, added auto compilation of CLI page
• @FrickTobias fixed a bug for memory info (#647 )
• fixed version checking for torch 2.0 release

- Python
Published by carsen-stringer almost 3 years ago

cellpose - cellpose v2.1.1

Implementing some minor fixes and updates * cellpose passes tests for python 3.9 and 3.10 across operating systems -- we still recommend python 3.8, particularly if you have issues with the install * added option to use Mac M1 chip (if installed with torch) with command line argument --gpu device mps * added option to run cellpose on a single file from the command like with --image_path as an alternative to --dir (#543) * added new suggestion mode to suggest best Cellpose 2 model * suppressed writing PNGs or outlines when no masks were found (#584) * added docs for all functions (#554) * fixed case sensitive image file detection (#555) * allow eval method of Cellpose and CellposeModel to take as input torch arrays (#557) * added the function io.add_model(/full/path/to/model) to add the model file to the hidden folder to use with GUI and CLI; can also access the function by running python -m cellpose --add_model /full/path/to/model (#558) * fixed issue with cursor not allowing resizing (#545 ) * added mac instructions (#523) * forced tiff saving for masks > 2**16 (#524) * added pyqt5 import statement for windows 8.1 compatibility

- Python
Published by carsen-stringer over 3 years ago

cellpose - cellpose v2.0.5

Small updates to Cellpose 2.0: * added new notebook for running Cellpose 2.0 * model zoo model that was last clicked is highlighted in GUI * number of ROIs in image is shown in the GUI * added stitch_threshold option to GUI * reenabled 3D annotations with "single stroke" option * added imagecodecs dependency * merged #494 and #495

- Python
Published by carsen-stringer almost 4 years ago

cellpose - cellpose v2.0.4

Cellpose 2.0 is here! Now you can train models in the loop with the Cellpose graphical interface. Upgrade to the latest version with pip install cellpose --upgrade

Check out the twitter thread and preprint for details.

Read how to use human-in-the-loop yourself in the gui docs, and check out the full human-in-the-loop video.

Also, read more about the model zoo and how to use user-trained models in the models docs.

- Python
Published by carsen-stringer almost 4 years ago

cellpose - Cellpose v1.0.2

fixes some bugs with the GUI and the colab and mintrainmasks

last stable release of Cellpose 1.0.

pip install cellpose==1.0.2

- Python
Published by carsen-stringer almost 4 years ago

cellpose - cellpose v1.0

stable release * fixes some bugs in 0.8 with plotting and flow visualization in GUI * merges pull request for stitchthreshold bug (#390) * allows images with empty masks, depending on new flag `--mintrainmasksthat is set to 5 by default. if fewer thanmintrainmasksin an image it is not used * resample set to defaultTrueas in older releases * model reloading per image turned off in CLI and class so that model is faster * optionalmasks.tifloading in GUI * warning for user if masks are to be saved innp.uint32`

- Python
Published by carsen-stringer about 4 years ago

cellpose - cellpose v0.8.0

• now install omnipose with pip! pip install omnipose
• added PR #416 which removed global logging settings, now turn on logging in a notebook with from cellpose.io import logger_setup; logger,log_file=logger_setup(), and from the command line with --verbose
• added support for >2^16 masks with np.uint32, if there are <2^16 masks then the masks are returned as np.uint16 still
• fixed bug with torch.long on windows

- Python
Published by carsen-stringer about 4 years ago

cellpose - cellpose | omnipose v0.7.3

Fixes to the following bugs: - --no_npy inverted settings - omni not passed through to remove_bad_flow_masks

- Python
Published by kevinjohncutler over 4 years ago

cellpose - Omnipose (cellpose v0.7.2)

Introducing Omnipose, a collaboration between the Stringer, Wiggins, and Mougous labs written by @kevinjohncutler. Read more about it in our preprint and on the Omnipose README. Important new features are: - cyto2_omni model for slight improvement over the 'cyto2' Cellpose model - bact_omni model for bacteria phase contrast segmentation (huge improvement over Cellpose models trained on bacteria, which you can demo with the bact model) - omni option to use Omnipose mask reconstruction with your Cellpose model to help reduce over-segmentation (off by default) - cluster option to force DBSCAN clustering in Omnipose mask reconstruction. This is off by default and turned on automatically when the average cell diameter is less than diam_threshold. Note theat scikit-learn is necessary for DBSCAN, and a CLI prompt will ask you to download it when you run --omni.

Several saving options have been included as well: - in_folders saves outputs into separate folders named masks, outlines, etc. (off by default) - dir_above saves output in the directory above the image directory (useful to have images next to masks etc.) (off by default) - save_txt turns on ImageJ outline saving (now off by default) - save_ncolor uses @kevinjohncutler's N-color algorithm to save masks with repeating but non-touching integers (typically 4 or fewer, 5 or 6 when necessary), which allows segmentations of thousands of cells to be presented without as many colors (which can become very hard to distinguish otherwise). Use in combination with a color map to visualize output.

Several bug fixes and pull requests are included in this release as well.

- Python
Published by kevinjohncutler over 4 years ago

cellpose - cellpose v0.6.1

fixes bugs with

• 2D resizing of flows
• training with CUDA in torch
• __main__.py relative imports -> absolute imports

- Python
Published by carsen-stringer about 5 years ago

cellpose - cellpose v0.6

Pytorch is now the default deep neural network software for cellpose. Mxnet will still be supported. To install mxnet (CPU), run pip install mxnet-mkl. To use mxnet in a notebook, declare torch=False when creating a model, e.g. model = models.Cellpose(torch=False). To use mxnet on the command line, add the flag --mxnet, e.g. python -m cellpose --dir ~/images/ --mxnet. The pytorch implementation is 20% faster than the mxnet implementation when running on the GPU and 20% slower when running on the CPU.

Dynamics are computed using bilinear interpolation by default instead of nearest neighbor interpolation. Set interp=False in model.eval to turn off. The bilinear interpolation will be slightly slower on the CPU, but it is faster than nearest neighbor if using torch and the GPU is enabled.

- Python
Published by carsen-stringer about 5 years ago

cellpose - cellpose v0.5

• sped up 3D segmentation by reducing padding
• tile_overlap as a parameter
• fixed bug with batch_size in CLI

- Python
Published by carsen-stringer about 5 years ago

cellpose - cellpose v0.1.0.1

• automated testing implemented
• dynamics are run at rescaled size (will be faster for images with cells larger than 30 pixels in diameter)
• pyinstaller binaries created from this release

- Python
Published by carsen-stringer over 5 years ago

cellpose - cellpose v0.0.3.1

removed matplotlib dependencies and added fast_mode that does not do test-time augmentations or average over 4 networks' outputs

- Python
Published by carsen-stringer over 5 years ago

cellpose - 3D segmentation + docs

• 3D segmentation enabled
• run and train at command line options debugged
• docs added

- Python
Published by carsen-stringer almost 6 years ago

cellpose - adding new features and fixing bugs

on pypi.

• fixed bug with running in notebook with single channel files

added - 3D labelling - command line interface for running in batch - color inversion option in gui - user can input diameter in pixels instead of "rescale"

- Python
Published by carsen-stringer almost 6 years ago