Overview

It’s highly recommended to start with prototyping with a subset of images in a jupyter notebook, then converting to a script for full scale analysis. This is a brief overview of what the workflow of using PhenoTypic would look like. For a more detailed explanation go to the tutorials. A simple workflow for PhenoTypic using a pre-built pipline would look like:

1. Tune pipeline and parameters

import phenotypic as pht
from phenotypic.prefab import HeavyWatershedPipeline

# Assuming 96-array format
image = pht.GridImage.imread("path/to/my/image.jpg", nrows=8, ncols=12)
fig, ax = image.show()
fig.show()

# This is the start of the parameter tuning step
# Create pipeline object. Class params are the settings for the pipeline
pipe = HeavyWatershedPipeline()
new_image_one = pipe.apply(image)

# GridImage.show_overlay() shows the objects detected in the image
overlay_fig, overlay_ax = new_image_one.show_overlay()
overlay_fig.show()

# If the default settings don't work for your images continue tuning
pipe_two = HeavyWatershedPipeline(border_remover_size=50)
new_image_two = pipe_two.apply(image)
overlay_two_fig, overlay_two_ax = new_image_two.show_overlay()
overlay_two_fig.show()

# Once the overlay detection looks like how you want, then you
#   can measure the objects in the image
meas = pipe_two.measure(new_image_two)

# export
overlay_two_fig.savefig("overlay.png")
meas.to_csv("measurements.csv")

2. Deploy as .py script

The following is a basic script for using PhenoTypic on a set of images. This can take a long time however. See the Getting Started tutorial for parallel processing.

from pathlib import Path
import phenotypic as pht
from phenotypic.prefab import HeavyWatershedPipeline

DIRPATH_IMAGES = Path("path/to/images")
IMAGE_SUFFIX = ".jpg" # This is the target files we're trying to import

filepaths = [
    filepath for filepath in DIRPATH_IMAGES.iterdir()
    if (filepath.is_file())                     # Ensures only files are read
        and (filepath.suffix == IMAGE_SUFFIX)   # Ensures only jpegs are read
]

# Create pipeline object.
pipe = HeavyWatershedPipeline(border_remover_size=50)

meas = []
for impath in filepaths:

    # Create a GridImage for a 96-well format plate
    image = pht.GridImage.imread(impath, nrows=8, ncols=12)
    current_meas = pipe.apply_and_measure(image, inplace=True)
    meas.append(current_meas)

meas = pd.concat(meas, axis=0)