Voronoi-Otsu labeling + instance segmentation code overhaul (#34)

* Instance segmentation refactor + Voronoi-Otsu - Improved code for instance segmentation - Added Voronoi-Otsu labeling from pyclesperanto TODO : credits for labeling * Disabled small removal in Voronoi-Otsu * Added new docs for instance seg * Docs + UI update - Updated welcome/README - Changed step for DoubleCounter * Update requirements.txt Fix typo * isort * Fix tests * Fixed parental issues and instance seg widget init - Fixed widgets parents that were incorrectly init - Improve use of instance seg. method classes and init * Fix inference * Added labeling tools + UI tweaks - Added tools from MLCourse to evaluate labels and auto-correct them - Instance seg benchmark notebook - Tweaked utils UI to scale according to Viewer size Co-Authored-By: gityves <[email protected]> * Testing instance methods Co-Authored-By: gityves <[email protected]> * Many fixes - Fixed monai reqs - Added custom functions for label checking - Fixed return type of voronoi_otsu and utils.resize - black * black * Complete instance method evaluation * Added pre-commit hooks * Enfore pre-commit style * Update .gitignore * Version bump * Updated project files * Fixed missing parent error * Fixed wrong value in instance sliders * Removing dask-image * Fixed erroneous dtype conversion * Update test_plugin_utils.py * Temporary test action patch * Update plugin_convert.py * Update tox.ini Added pocl for testing on GH Actions * Update tox.ini * Found existing pocl * Updated utils test to avoid Voronoi-Otsu VO is missing CL runtime * Relabeling tests * Run full suite of pre-commit hooks * Enforce style * Instance segmentation refactor + Voronoi-Otsu - Improved code for instance segmentation - Added Voronoi-Otsu labeling from pyclesperanto TODO : credits for labeling * Disabled small removal in Voronoi-Otsu * Added new docs for instance seg * isort * Fix tests * Fixed parental issues and instance seg widget init - Fixed widgets parents that were incorrectly init - Improve use of instance seg. method classes and init * Fix inference * Added labeling tools + UI tweaks - Added tools from MLCourse to evaluate labels and auto-correct them - Instance seg benchmark notebook - Tweaked utils UI to scale according to Viewer size Co-Authored-By: gityves <[email protected]> * Testing instance methods Co-Authored-By: gityves <[email protected]> * Many fixes - Fixed monai reqs - Added custom functions for label checking - Fixed return type of voronoi_otsu and utils.resize - black * black * Complete instance method evaluation * Enfore pre-commit style * Removing dask-image * Fixed erroneous dtype conversion * Update test_plugin_utils.py * Update tox.ini * Added new pre-commit hooks * Run full suite of pre-commit hooks * Enforce style * Documentation update, crop contrast fix * Update plugin_model_inference.py * Updated hooks --------- Co-authored-by: gityves <[email protected]>
AdaptiveMotorControlLab · Jun 11, 2023 · e00806c · e00806c
1 parent 89e71dc
commit e00806c
Show file tree

Hide file tree

Showing 43 changed files with 2,857 additions and 618 deletions.
diff --git a/.github/workflows/test_and_deploy.yml b/.github/workflows/test_and_deploy.yml
@@ -8,12 +8,14 @@ on:
     branches:
       - main
       - npe2
+      - cy/voronoi-otsu
     tags:
       - "v*" # Push events to matching v*, i.e. v1.0, v20.15.10
   pull_request:
     branches:
       - main
       - npe2
+      - cy/voronoi-otsu
   workflow_dispatch:
 
 jobs:

diff --git a/.gitignore b/.gitignore
@@ -105,3 +105,8 @@ notebooks/full_plot.html
 *.csv
 *.png
 *.prof
+
+#include test data
+!napari_cellseg3d/_tests/res/test.tif
+!napari_cellseg3d/_tests/res/test.png
+!napari_cellseg3d/_tests/res/test_labels.tif
diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
@@ -5,11 +5,14 @@ repos:
 #      - id: check-docstring-first
       - id: end-of-file-fixer
       - id: trailing-whitespace
-  - repo: https://github.com/pycqa/isort
-    rev: 5.12.0
-    hooks:
-      - id: isort
-        args: ["--profile", "black", --line-length=79]
+      - id: check-yaml
+      - id: check-added-large-files
+      - id: check-toml
+#  - repo: https://github.com/pycqa/isort
+#    rev: 5.12.0
+#    hooks:
+#      - id: isort
+#        args: ["--profile", "black", --line-length=79]
   - repo: https://github.com/charliermarsh/ruff-pre-commit
     # Ruff version.
     rev: 'v0.0.262'

diff --git a/README.md b/README.md
@@ -151,8 +151,9 @@ Distributed under the terms of the [MIT] license.
 
 ## Acknowledgements
 
-This plugin was developed by Cyril Achard, Maxime Vidal, Mackenzie Mathis. This work was funded, in part, from the Wyss Center to the [Mathis Laboratory of Adaptive Motor Control](https://www.mackenziemathislab.org/).
-
+This plugin was developed by Cyril Achard, Maxime Vidal, Mackenzie Mathis.
+This work was funded, in part, from the Wyss Center to the [Mathis Laboratory of Adaptive Motor Control](https://www.mackenziemathislab.org/).
+Please refer to the documentation for full acknowledgements.
 
 ## Plugin base
 This [napari] plugin was generated with [Cookiecutter] using [@napari]'s [cookiecutter-napari-plugin] template.

diff --git a/docs/res/code/model_instance_seg.rst b/docs/res/code/model_instance_seg.rst
@@ -1,6 +1,29 @@
 model_instance_seg.py
 ===========================================
 
+Classes
+-------------
+
+InstanceMethod
+**************************************
+.. autoclass:: napari_cellseg3d.code_models.model_instance_seg::InstanceMethod
+    :members: __init__
+
+ConnectedComponents
+**************************************
+.. autoclass:: napari_cellseg3d.code_models.model_instance_seg::ConnectedComponents
+    :members: __init__
+
+Watershed
+**************************************
+.. autoclass:: napari_cellseg3d.code_models.model_instance_seg::Watershed
+    :members: __init__
+
+VoronoiOtsu
+**************************************
+.. autoclass:: napari_cellseg3d.code_models.model_instance_seg::VoronoiOtsu
+    :members: __init__
+
 
 Functions
 -------------

diff --git a/docs/res/code/plugin_convert.rst b/docs/res/code/plugin_convert.rst
@@ -19,11 +19,6 @@ ToSemanticUtils
 .. autoclass:: napari_cellseg3d.code_plugins.plugin_convert::ToSemanticUtils
     :members: __init__
 
-InstanceWidgets
-**********************************
-.. autoclass:: napari_cellseg3d.code_plugins.plugin_convert::InstanceWidgets
-    :members: __init__, run_method
-
 ToInstanceUtils
 **********************************
 .. autoclass:: napari_cellseg3d.code_plugins.plugin_convert::ToInstanceUtils

diff --git a/docs/res/guides/cropping_module_guide.rst b/docs/res/guides/cropping_module_guide.rst
@@ -33,9 +33,9 @@ If you'd like to change the size of the volume, change the parameters as previou
 
 Creating new layers
 ---------------------------------
-To "zoom in" your volume, you can use the "Create new layers" checkbox to make a new layer not controlled by the plugin next
-time you hit Start. This way, you can first select your region of interest by using the tool as described above,
-the enable the option, select the cropped layer, and define a smaller crop size to have easier access to your region of interest.
+To "zoom in" your volume, you can use the "Create new layers" checkbox to make a new cropping layer controlled by the sliders
+next time you hit Start. This way, you can first select your region of interest by using the tool as described above,
+then enable the option, select the cropped region produced before as the input layer, and define a smaller crop size in order to crop within your region of interest.
 
 Interface & functionalities
 ---------------------------------------------------------------

diff --git a/docs/res/guides/utils_module_guide.rst b/docs/res/guides/utils_module_guide.rst
@@ -4,13 +4,21 @@ Label conversion utility guide
 ==================================
 
 This utility will let you convert labels to various different formats.
+
 You will have to specify the results directory for saving; afterwards you can run each action on a folder or on the currently selected layer.
 
 You can :
 
+* Crop 3D volumes :
+    Please refer to :ref:`cropping_module_guide` for a guide on using the cropping utility.
+
 * Convert to instance labels :
-    This will convert 0/1 semantic labels to instance label, with a unique ID for each object using the watershed method.
+    This will convert 0/1 semantic labels to instance label, with a unique ID for each object.
+    The available methods for this are :
 
+    * Connected components : simple method that will assign a unique ID to each connected component. Does not work well for touching objects (objects will often be fused), works for anisotropic volumes.
+    * Watershed : method based on topographic maps. Works well for touching objects and anisotropic volumes; touching objects may be fused.
+    * Voronoi-Otsu : method based on Voronoi diagrams. Works well for touching objects but only for isotropic volumes.
 * Convert to semantic labels :
     This will convert instance labels with unique IDs per object into 0/1 semantic labels, for example for training.
 

diff --git a/docs/res/welcome.rst b/docs/res/welcome.rst
@@ -38,22 +38,28 @@ You can install `napari-cellseg3d` via [pip]:
 
   ``pip install napari-cellseg3d``
 
-  For local installation, please run:
+For local installation after cloning, please run in the CellSeg3D folder:
 
   ``pip install -e .``
 
 Requirements
 --------------------------------------------
 
+.. note::
+    A **CUDA-capable GPU** is not needed but **very strongly recommended**, especially for training and possibly inference.
+
 .. important::
-    A **CUDA-capable GPU** is not needed but **very strongly recommended**, especially for training.
+    This package requires you have napari installed with PyQt5 or PySide2 first.
+    If you do not have a Qt backend you can use :
 
-This package requires you have napari installed first.
+        ``pip install napari-cellseg3d[all]``
+    to install PyQt5 by default.
 
-It also depends on PyTorch and some optional dependencies of MONAI. These come in the pip package above, but if
+It also depends on PyTorch and some optional dependencies of MONAI. These come in the pip package as requirements, but if
 you need further assistance see below.
 
 * For help with PyTorch, please see `PyTorch's website`_ for installation instructions, with or without CUDA depending on your hardware.
+  Depending on your setup, you might wish to install torch first.
 
 * If you get errors from MONAI regarding missing readers, please see `MONAI's optional dependencies`_ page for instructions on getting the readers required by your images.
 
@@ -70,14 +76,13 @@ To use the plugin, please run:
 
 Then go into Plugins > napari-cellseg3d, and choose which tool to use:
 
-
 - **Review**: This module allows you to review your labels, from predictions or manual labeling, and correct them if needed. It then saves the status of each file in a csv, for easier monitoring
 - **Inference**: This module allows you to use pre-trained segmentation algorithms on volumes to automatically label cells
 - **Training**:  This module allows you to train segmentation algorithms from labeled volumes
 - **Utilities**: This module allows you to use several utilities, e.g. to crop your volumes and labels, compute prediction scores or convert labels
 - **Help/About...** : Quick access to version info, Github page and docs
 
-See above for links to detailed guides regarding the usage of the modules.
+See the documentation for links to detailed guides regarding the usage of the modules.
 
 Acknowledgments & References
 ---------------------------------------------
@@ -90,24 +95,29 @@ We also provide a model that was trained in-house on mesoSPIM nuclei data in col
 
 This plugin mainly uses the following libraries and software:
 
-* `napari website`_
+* `napari`_
 
-* `PyTorch website`_
+* `PyTorch`_
 
-* `MONAI project website`_ (various models used here are credited `on their website`_)
+* `MONAI project`_ (various models used here are credited `on their website`_)
 
+* `pyclEsperanto`_ (for the Voronoi Otsu labeling) by Robert Haase
+
+* A custom re-implementation of the `WNet model`_ by Xia and Kulis [#]_
 
 .. _Mathis Laboratory of Adaptive Motor Control: http://www.mackenziemathislab.org/
 .. _Wyss Center: https://wysscenter.ch/
 .. _TRAILMAP project on GitHub: https://github.com/AlbertPun/TRAILMAP
-.. _napari website: https://napari.org/
-.. _PyTorch website: https://pytorch.org/
-.. _MONAI project website: https://monai.io/
+.. _napari: https://napari.org/
+.. _PyTorch: https://pytorch.org/
+.. _MONAI project: https://monai.io/
 .. _on their website: https://docs.monai.io/en/stable/networks.html#nets
-
+.. _pyclEsperanto: https://github.com/clEsperanto/pyclesperanto_prototype
+.. _WNet model: https://arxiv.org/abs/1711.08506
 
 .. rubric:: References
 
 .. [#] Mapping mesoscale axonal projections in the mouse brain using a 3D convolutional network, Friedmann et al., 2020 ( https://pnas.org/cgi/doi/10.1073/pnas.1918465117 )
 .. [#] The mesoSPIM initiative: open-source light-sheet microscopes for imaging cleared tissue, Voigt et al., 2019 ( https://doi.org/10.1038/s41592-019-0554-0 )
 .. [#] MONAI Project website ( https://monai.io/ )
+.. [#] W-Net: A Deep Model for Fully Unsupervised Image Segmentation, Xia and Kulis, 2018 ( https://arxiv.org/abs/1711.08506 )
diff --git a/napari_cellseg3d/_tests/pytest.ini b/napari_cellseg3d/_tests/pytest.ini
@@ -0,0 +1,2 @@
+[pytest]
+qt_api=pyqt5
diff --git a/napari_cellseg3d/_tests/res/test_labels.tif b/napari_cellseg3d/_tests/res/test_labels.tif
diff --git a/napari_cellseg3d/_tests/test_labels_correction.py b/napari_cellseg3d/_tests/test_labels_correction.py
@@ -0,0 +1,52 @@
+from pathlib import Path
+
+import numpy as np
+from tifffile import imread
+
+from napari_cellseg3d.dev_scripts import artefact_labeling as al
+from napari_cellseg3d.dev_scripts import correct_labels as cl
+from napari_cellseg3d.dev_scripts import evaluate_labels as el
+
+res_folder = Path(__file__).resolve().parent / "res"
+image_path = res_folder / "test.tif"
+image = imread(str(image_path))
+
+labels_path = res_folder / "test_labels.tif"
+labels = imread(str(labels_path))  # .astype(np.int32)
+
+
+def test_artefact_labeling():
+    output_path = str(res_folder / "test_artifacts.tif")
+    al.create_artefact_labels(image, labels, output_path=output_path)
+    assert Path(output_path).is_file()
+
+
+def test_artefact_labeling_utils():
+    crop_test = al.crop_image(image)
+    assert isinstance(crop_test, np.ndarray)
+    output_path = str(res_folder / "test_cropped.tif")
+    al.crop_image_path(image, path_image_out=output_path)
+    assert Path(output_path).is_file()
+
+
+def test_correct_labels():
+    output_path = res_folder / "test_correct"
+    output_path.mkdir(exist_ok=True, parents=True)
+    cl.relabel_non_unique_i(
+        labels, str(output_path / "corrected.tif"), go_fast=True
+    )
+
+
+def test_relabel(make_napari_viewer):
+    viewer = make_napari_viewer()
+    cl.relabel(
+        str(image_path),
+        str(labels_path),
+        go_fast=True,
+        viewer=viewer,
+        test=True,
+    )
+
+
+def test_evaluate_model_performance():
+    el.evaluate_model_performance(labels, labels, print_details=True)
diff --git a/napari_cellseg3d/_tests/test_plugin_utils.py b/napari_cellseg3d/_tests/test_plugin_utils.py
@@ -1,16 +1,32 @@
-from napari_cellseg3d.code_plugins.plugin_utilities import Utilities
+from pathlib import Path
+
+import numpy as np
+from tifffile import imread
+
 from napari_cellseg3d.code_plugins.plugin_utilities import (
     UTILITIES_WIDGETS,
+    Utilities,
 )
 
 
 def test_utils_plugin(make_napari_viewer):
     view = make_napari_viewer()
     widget = Utilities(view)
 
+    im_path = str(Path(__file__).resolve().parent / "res/test.tif")
+    image = imread(im_path)
+    view.add_image(image)
+    view.add_labels(image.astype(np.uint8))
+
     view.window.add_dock_widget(widget)
     for i, utils_name in enumerate(UTILITIES_WIDGETS.keys()):
         widget.utils_choice.setCurrentIndex(i)
         assert isinstance(
             widget.utils_widgets[i], UTILITIES_WIDGETS[utils_name]
         )
+        if utils_name == "Convert to instance labels":
+            # to avoid issues with Voronoi-Otsu missing runtime
+            menu = widget.utils_widgets[i].instance_widgets.method_choice
+            menu.setCurrentIndex(menu.currentIndex() + 1)
+
+        widget.utils_widgets[i]._start()
diff --git a/napari_cellseg3d/_tests/test_weight_download.py b/napari_cellseg3d/_tests/test_weight_download.py
@@ -1,5 +1,7 @@
-from napari_cellseg3d.code_models.model_workers import WEIGHTS_DIR
-from napari_cellseg3d.code_models.model_workers import WeightsDownloader
+from napari_cellseg3d.code_models.model_workers import (
+    WEIGHTS_DIR,
+    WeightsDownloader,
+)
 
 
 # DISABLED, causes GitHub actions to freeze

diff --git a/napari_cellseg3d/code_models/model_framework.py b/napari_cellseg3d/code_models/model_framework.py
@@ -5,13 +5,11 @@
 import torch
 
 # Qt
-from qtpy.QtWidgets import QProgressBar
-from qtpy.QtWidgets import QSizePolicy
+from qtpy.QtWidgets import QProgressBar, QSizePolicy
 
 # local
-from napari_cellseg3d import config
+from napari_cellseg3d import config, utils
 from napari_cellseg3d import interface as ui
-from napari_cellseg3d import utils
 from napari_cellseg3d.code_plugins.plugin_base import BasePluginFolder
 
 warnings.formatwarning = utils.format_Warning
@@ -286,11 +284,10 @@ def _load_weights_path(self):
         )
         if file[0] == self._default_weights_folder:
             return
-        if file is not None:
-            if file[0] != "":
-                self.weights_config.path = file[0]
-                self.weights_filewidget.text_field.setText(file[0])
-                self._default_weights_folder = str(Path(file[0]).parent)
+        if file is not None and file[0] != "":
+            self.weights_config.path = file[0]
+            self.weights_filewidget.text_field.setText(file[0])
+            self._default_weights_folder = str(Path(file[0]).parent)
 
     @staticmethod
     def get_device(show=True):