Combining algorithms

Combining algorithms#

Combining algorithms allows users to access multiple algorithms from a single interface, such as the “Algorithm” dropdown in Napari (or the equivalent dropdown in QuPath).

Using sk.combine#

Let’s consider two separately defined algorithms:

import imaging_server_kit as sk

@sk.algorithm(
    name="threshold", 
    description="Segment a grayscale image based on an intensity threshold."
)
def threshold_algo(image, threshold=128):
    mask = image > threshold
    return sk.Mask(mask, name="Binary mask")

@sk.algorithm(
    name="foreground", 
    description="Compute the fraction of positive pixels in a binary mask."
)
def foreground_fract(mask):
    fract = mask.sum() / mask.size
    return sk.Float(fract, name="Foreground fraction")

You can combine these algorithms into a single collection using sk.combine:

multi_algo = sk.combine([threshold_algo, foreground_fract], name="segmentation-pipeline")

# Test the combined algorithms in Napari
sk.to_napari(multi_algo)

Using this technique, the two algorithms become available from the “Algorithm” dropdown in Napari.

Summary#

  • Use sk.combine to create a collection of multiple algorithms and use them in a single interface.

Next steps#

In the next section, we will explore the concept of live updates in Imaging Server Kit.

Contents