Quick start guideΒΆ

Here we illustrate the use of jicbioimage to segment an image of Greek coins from Pompeii.

First we load the image from scikit-image as a numpy array.

>>> import skimage.data
>>> array = skimage.data.coins()

We then create a jicbioimage.core.image.Image instance from the array.

>>> from jicbioimage.core.image import Image
>>> image = Image.from_array(array)

If using IPython qtconsole/notebook the image can be viewed directly in the interpreter.

>>> image
Coins.

We can now threshold the image, for example using Otsu’s method.

>>> from jicbioimage.transform import threshold_otsu
>>> image = threshold_otsu(image)
>>> image
Coins thresholded.

Let us segment the thresholded image into connected components.

>>> from jicbioimage.segment import connected_components
>>> segmentation = connected_components(image, background=0)
>>> segmentation
Coins segmented.

The jicbioimage.segment.connected_components() function returns an instance of the jicbioimage.segment.SegmentedImage class, which provides access to segmented regions of interest as jicbioimage.segment.Region instances.

Finally, let us write a couple of functions to create an augmented reality image.

>>> import numpy as np
>>> from jicbioimage.illustrate import AnnotatedImage
>>> def text_position(region):
...     "Return x, y coordinates of text position."
...     ys, xs = region.index_arrays
...     y = np.min(ys) - 5
...     x = np.mean(xs, dtype=int)
...     return (y, x)
...
>>> def augment_image(image, segmentation):
...     "Return an augmented image."
...     augmented = AnnotatedImage.from_grayscale(image)
...     for i in segmentation.identifiers:
...         region = segmentation.region_by_identifier(i)
...         if region.area > 300 and region.area < 5000:
...             augmented.mask_region(region.convex_hull.border)
...             pos = text_position(region.convex_hull)
...             text = "{}px".format(region.convex_hull.area)
...             augmented.text_at(text, pos, center=True, antialias=False)
...     return augmented
...
>>> augmented = augment_image(array, segmentation)
>>> augmented
Coins augmented.