add course materials

.gitignore

+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+*.pyc
+*.~*.py
+*.py.swm
+*.py.swn
+*.py.swp
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+.hypothesis/
+.pytest_cache/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# pyenv
+.python-version
+
+# celery beat schedule file
+celerybeat-schedule
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+
+.DS_Store
+
+# personal stuff
+Part1Unet/models
+Part1Unet/logs
+Part1Unet/unet_model.png
+Part2Cellpose/models

Part1Unet/README.md

+# Unet
+
+Install unet env in conda
+
+```
+setconda
+conda env create -f unet_env.yml
+```
+
+Launch the jupyter notebook
+
+```
+setconda
+conda activate qlife25-unet
+setenv LD_LIBRARY_PATH "${CONDA_PREFIX}/lib:${LD_LIBRARY_PATH}"
+jupyter notebook
+```

Part1Unet/unet_env.yml

+name: qlife25-unet
+channels:
+  - conda-forge
+dependencies:
+  - python=3.10
+  - numpy=1.23.5
+  - cudatoolkit=11.2
+  - cudnn=8.1.0
+  - pip:
+      - tensorflow-gpu==2.10.0
+      - keras==2.10.0
+      - notebook==6.1.5
+      - scikit-image==0.19.3
+      - matplotlib==3.7.1
+      - pydot==3.0.4

Part1Unet/utils.py

+"""
+data_loader.py
+
+This script contains functions for handling various data processing tasks for the 
+Cell-Segmentation-UNet project. It includes functions for unzipping datasets, retrieving data paths, 
+preprocessing images and masks, and loading datasets into a suitable format for training and testing.
+
+Functions:
+- unzip_to_original_subfolders: Unzips all files within a directory to their respective subfolders.
+- get_data_paths: Retrieves paths to folders named by unique IDs within a specified directory.
+- preprocess_image: Processes an image file to fit the input requirements of the model.
+- preprocess_mask: Processes mask files associated with each image for segmentation tasks.
+- load_dataset: Loads and preprocesses the dataset from a specified directory for model training.
+
+Author: Mahdi Habibi
+"""
+
+import os
+import random
+import numpy as np
+import matplotlib.pyplot as plt
+from skimage.io import imread
+from skimage.transform import resize
+
+def unzip_to_original_subfolders(root_directory):
+    """
+    Unzip all files within a directory to their respective subfolders.
+    
+    Parameters:
+    - root_directory (str): Path to the root directory containing zip files.
+    """
+    # Walking through the root directory to find all files
+    for dirpath, _, filenames in os.walk(root_directory):
+        for filename in filenames:
+            file_path = os.path.join(dirpath, filename)
+            # Extract the file name without the extension to create a subfolder
+            folder_name = os.path.splitext(filename)[0]
+            destination_folder = os.path.join(dirpath, folder_name)
+            # Unzipping to the created subfolder
+            os.system(f'unzip -q {file_path} -d {destination_folder}')
+            # remove the zip file after extracting
+            os.remove(file_path)
+            
+def get_data_paths(path):
+    """
+    Retrieve paths to the folders named by IDs within the specified training path.
+    
+    Parameters:
+    - path (str): Path to the data directory.
+    
+    Returns:
+    - list: A list of paths to the individual data folders.
+    """
+    return next(os.walk(path))[1]
+
+def preprocess_image(id_, path, img_height, img_width, img_channels):
+    """
+    Preprocess an image file to fit the input requirements of the model.
+    
+    Parameters:
+    - id_ (str): Unique identifier for the image.
+    - path (str): Path to the data directory.
+    - img_height (int): Desired image height.
+    - img_width (int): Desired image width.
+    - img_channels (int): Number of image channels.
+    
+    Returns:
+    - ndarray: Processed image array.
+    """
+    img_file_path = os.path.join(path, id_, "images", id_ + ".png")
+    img = imread(img_file_path)[:, :, :img_channels]
+    img = resize(img, (img_height, img_width), mode='constant', preserve_range=True)
+    return img
+
+def preprocess_mask(id_, path, img_height, img_width):
+    """
+    Process mask files associated with each image for segmentation tasks.
+    
+    Parameters:
+    - id_ (str): Unique identifier for the image.
+    - path (str): Path to the data directory.
+    - img_height (int): Desired mask height.
+    - img_width (int): Desired mask width.
+    
+    Returns:
+    - ndarray: Combined mask array for the given image.
+    """
+    mask_file_dir = os.path.join(path, id_, "masks")
+    all_masks = os.listdir(mask_file_dir)
+    mask = np.zeros((img_height, img_width, 1), dtype=np.bool)
+    for mask_file in all_masks:
+        mask_path = os.path.join(mask_file_dir, mask_file)
+        mask_ = imread(mask_path)
+        mask_ = np.expand_dims(resize(mask_, (img_height, img_width), mode='constant', preserve_range=True), axis=-1)
+        mask = np.maximum(mask, mask_)
+    return mask
+
+def load_train_dataset(train_path, img_height, img_width, img_channels):
+    """
+    Loads and preprocesses the dataset from the specified directory for model training.
+    
+    Parameters:
+    - train_path (str): Path to the training data directory.
+    - img_height (int): Desired image height.
+    - img_width (int): Desired image width.
+    - img_channels (int): Number of image channels.
+    
+    Returns:
+    - tuple: Tuple containing arrays for training images and corresponding masks.
+    """
+    train_ids = get_data_paths(train_path)
+
+    X_train = np.zeros((len(train_ids), img_height, img_width, img_channels), dtype=np.uint8)
+    Y_train = np.zeros((len(train_ids), img_height, img_width, 1), dtype=np.bool)
+
+    for n, id_ in enumerate(train_ids):
+        X_train[n] = preprocess_image(id_, train_path, img_height, img_width, img_channels)
+        Y_train[n] = preprocess_mask(id_, train_path, img_height, img_width)
+
+    return X_train, Y_train
+
+def load_test_dataset(test_path, img_height, img_width, img_channels):
+    """
+    Loads and preprocesses the dataset from the specified directory for model training.
+    
+    Parameters:
+    - test_path (str): Path to the test data directory.
+    - img_height (int): Desired image height.
+    - img_width (int): Desired image width.
+    - img_channels (int): Number of image channels.
+    
+    Returns:
+    - tuple: Tuple containing arrays for training images and corresponding masks.
+    """
+    test_ids = get_data_paths(test_path)
+
+    X_test = np.zeros((len(test_ids), img_height, img_width, img_channels), dtype=np.uint8)
+
+    for n, id_ in enumerate(test_ids):
+        X_test[n] = preprocess_image(id_, test_path, img_height, img_width, img_channels)
+
+    return X_test
+    
+def display_sample_grid(X, Y, num_samples=3):
+    """Display a grid of microscopy images and their associated masks."""
+    fig, axs = plt.subplots(2, num_samples, figsize=(12, 8))  
+
+    for i in range(num_samples):
+        idx = random.randint(0, len(X) - 1)
+        
+        # Display original image
+        axs[0, i].imshow(X[idx])
+        axs[0, i].set_title(f'Original Image {idx}')
+        axs[0, i].axis('off')  # Turn off axis
+
+        # Display corresponding mask
+        axs[1, i].imshow(np.squeeze(Y[idx]))
+        axs[1, i].set_title(f'Mask Image {idx}')
+        axs[1, i].axis('off')  # Turn off axis
+
+    plt.tight_layout()
+    plt.show()
+
+def display_train_prediction(X_data, Y_data, predictions, predictions_thr,_ix=None):
+    """
+    Display a random sample consisting of the original image, ground truth mask, and predicted mask.
+
+    Parameters:
+    - X_data: Array of original images.
+    - Y_data: Array of ground truth masks.
+    - predictions: Array of predicted masks by the model.
+
+    The function randomly selects an image from the dataset, displays the original image,
+    the actual mask (ground truth), and the mask predicted by the model. This visual comparison
+    helps in understanding the model's segmentation accuracy on a qualitative level.
+    """
+    # Randomly select an index from the dataset
+    if _ix is None:
+        ix = random.randint(0, len(predictions) - 1)
+    else:
+        ix = _ix
+
+    # Plot original image, ground truth mask, and predicted mask
+    fig, ax = plt.subplots(1, 4, figsize=(16, 8))
+    ax[0].imshow(X_data[ix])
+    ax[0].title.set_text(f'Original Image {ix}')
+
+    ax[1].imshow(np.squeeze(Y_data[ix]))
+    ax[1].title.set_text('Ground Truth Mask')
+
+    ax[2].imshow(np.squeeze(predictions[ix]))
+    ax[2].title.set_text('Predicted Probability')
+
+    ax[3].imshow(np.squeeze(predictions_thr[ix]))
+    ax[3].title.set_text('Predicted Mask')
+
+    plt.tight_layout()
+    plt.show()
+
+def display_test_prediction(X_data, predictions, predictions_thr,_ix=None):
+    """
+    Display a random sample consisting of the original image, ground truth mask, and predicted mask.
+
+    Parameters:
+    - X_data: Array of original images.
+    - predictions: Array of predicted masks by the model.
+
+    The function randomly selects an image from the dataset, displays the original image,
+    the actual mask (ground truth), and the mask predicted by the model. This visual comparison
+    helps in understanding the model's segmentation accuracy on a qualitative level.
+    """
+    # Randomly select an index from the dataset
+    if _ix is None:
+        ix = random.randint(0, len(predictions) - 1)
+    else:
+        ix = _ix
+
+    # Plot original image, ground truth mask, and predicted mask
+    fig, ax = plt.subplots(1, 3, figsize=(16, 8))
+    ax[0].imshow(X_data[ix])
+    ax[0].title.set_text(f'Original Image {ix}')
+
+    ax[1].imshow(np.squeeze(predictions[ix]))
+    ax[1].title.set_text('Predicted Probability')
+
+    ax[2].imshow(np.squeeze(predictions_thr[ix]))
+    ax[2].title.set_text('Predicted Mask')
+
+    plt.tight_layout()
+    plt.show()

Part2Cellpose/README.md

+# Cellpose
+
+Run the following commands to install cellpose env in conda
+
+```
+setconda
+conda create -n qlife25-cellpose python=3.10 -y
+conda activate qlife-cellpose
+conda install pytorch==1.12.0 cudatoolkit=11.3 -c pytorch
+pip uninstall torch
+pip install 'cellpose[gui]'
+pip install notebook==6.1.5 scikit-image matplotlib
+```
+
+Launch the jupyter notebook
+
+```
+setconda
+conda activate qlife25-cellpose
+jupyter notebook
+```
\ No newline at end of file