Pipeline examples

from easycv import Image, Pipeline
from easycv.transforms import Gradient, Blur, Rotate

For this example we will load an image using random().

img = Image.random()
img

Creating pipeline

To create a pipeline we simply call it’s constructor with a list of transforms.

pipeline = Pipeline([Gradient(), Blur()], name="test")
pipeline

Pipeline (test) with 2 transforms
    1: Gradient (axis=x, method=sobel, size=5)
    2: Blur (method=gaussian, size=auto, sigma=0, sigma_color=75, sigma_space=75, truncate=4)

img.apply(pipeline)

We can also create a pipeline with another pipeline inside

pipeline = Pipeline([pipeline, Rotate(degrees=20)],name="test")
pipeline

Pipeline (test) with 3 transforms
    1: Pipeline (test) with 2 transforms
    |    1: Gradient (axis=x, method=sobel, size=5)
    |    2: Blur (method=gaussian, size=auto, sigma=0, sigma_color=75, sigma_space=75, truncate=4)
    2: Rotate (degrees=20, scale=1, center=auto, original=True)

And it’s possible to add a transform to an existing pipeline

pipeline.add_transform(Rotate(degrees=-40))
pipeline

Pipeline (test) with 4 transforms
    1: Pipeline (test) with 2 transforms
    |    1: Gradient (axis=x, method=sobel, size=5)
    |    2: Blur (method=gaussian, size=auto, sigma=0, sigma_color=75, sigma_space=75, truncate=4)
    2: Rotate (degrees=20, scale=1, center=auto, original=True)
    3: Rotate (degrees=-40, scale=1, center=auto, original=True)

Number of transforms

You can check the number of transforms in a pipeline. This is the number of Transforms so nested pipelines count as their internal number of Transforms

pipeline.num_transforms()

4

Clear

You can also remove every transform in a pipeline

pipeline.clear()
pipeline.num_transforms()

0

Save

If you want you can save a pipeline for future use or sharing

pipeline.save(filename="pipeline")