In [1]:
import bia_bob
import pyclesperanto_prototype as cle
import stackview
In [2]:
import numpy as np # Already imported modules should not be imported again
from skimage.data import cells3d
# Extract the nuclei channel (assuming channel 1 is the nuclei channel)
nuclei_channel = cells3d()[:, 1, :, :]
nuclei_channel.shape # to verify the extraction
Out[2]:
(60, 256, 256)
In [3]:
# Apply Otsu's threshold to create a binary image
binary_nuclei = cle.threshold_otsu(nuclei_channel)
# Perform connected component labeling
labeled_nuclei = cle.connected_components_labeling_box(binary_nuclei)
# Display the segmented label map
stackview.insight(labeled_nuclei)
Out[3]:
|
|
In [4]:
import pandas as pd
from skimage.measure import regionprops_table
# Define properties to extract
properties = ['label', 'area', 'mean_intensity']
# Extract features
measurements = regionprops_table(np.asarray(labeled_nuclei), intensity_image=nuclei_channel, properties=properties)
# Store results in a DataFrame
df = pd.DataFrame(measurements)
df.head() # To display the first few rows of the extracted features
Out[4]:
| label | area | mean_intensity | |
|---|---|---|---|
| 0 | 1 | 27119.0 | 12862.000295 |
| 1 | 2 | 5.0 | 11143.800000 |
| 2 | 3 | 168.0 | 11268.273810 |
| 3 | 4 | 60.0 | 11758.683333 |
| 4 | 5 | 1.0 | 10527.000000 |
In [6]:
import stackview
stackview.clusterplot(df, labeled_nuclei, column_x="area", column_y="mean_intensity", image=nuclei_channel)
Out[6]:
VBox(children=(HBox(children=(VBox(children=(HBox(children=(VBox(children=(ImageWidget(height=256, width=256),…
In [ ]: